Skip to main content

Full text of "Analysis of Variability and Condition of Cavate Structures in Bandelier National Monument"

See other formats


INTERMOUNTAIN CULTURAL RESOURCES CENTER 



ANTHROPOLOGY PROGRAM 



OCUMENTS 



JAN 1 1995 

An Analysis of Variability and Condition 

of Cavate Structures in Bandelier ^National — J 

Monument 



H. WOLCOTT TOLL 




1939 



Intermountain Cultural Resources Center 



1986 



Q G-008?-p 



Professional Paper No. 53 



Intermountain Cultural Resources Center 
Professional Paper No. 53 



An Analysis of Variability and Condition of 
Cavate Structures in Bandelier National 

Monument 



By H. Wolcott Toll 

With contributions by 

Peter J. McKenna and June Crowder 



Contribution #3 of the Bandelier Archeological Survey 



Anthropology Program 

U.S. Department of the Interior 

National Park Service 

1995 



Mission 

As the Nation's principal conservation agency, the Department 
of the Interior has responsibility for most of our nationally-owned public 
lands and natural and cultural resources. This includes fostering wise use 
of our land and water resources, protecting our fish and wildlife, 
preserving the environmental and cultural values of our national parks 
and historical places, and providing for the enjoyment of life through 
outdoor recreation. The Department assesses our energy and mineral 
resources and works to assure that their development is in the best 
interests of all our people. The Department also promotes the goals of 
the Take Pride in America campaign by encouraging stewardship and 
citizen responsibility for the public lands and promoting citizen 
participation in their care. The Department also has a major 
responsibility for American Indian reservation communities and for 
people who live in Island Territories under U.S. Administration. 





CONTRIBUTIONS OF 
THE BANDELIER ARCHEOLOGICAL SURVEY 



1 M ATfflEN, FRANCES JOAN, CHARLIE R. STEEN, AND CRAIG D. ALLEN 

1993 The Pajarito Plateau: A Bibliography. Southwest Cultural Resources Center 

Professional Paper No. 49. Santa Fe. 

2 KOHLER, TIMOTHY A. (editor) 

1989 Bandelier Archaeological Excavation Project: Research Design and Summer 1988 
Sampling. Reports of Investigations No. 61. Department of Anthropology, Washington 
State University, Pullman. 

3 TOLL, H. WOLCOTT 

1995 An Analysis of Variability and Condition of Cavate Structures in Bandelier National 

Monument. Intermountain Cultural Resources Center Professional Paper No. 53. Santa 
Fe. 

4 MATHIEN, FRANCES JOAN 

1991 Glimpses into the History of the 1908 Fieldwork at Yapashi, Bandelier National 
Monument. In Puebloan Past and Present: Papers in Honor of Stewart Peckham, edited 
by Meliha S. Duran and David T. Kirkpatrick, pp. 121-132. Papers of the Archeological 
Society of New Mexico: 17. Albuquerque. 

5 KOHLER, TIMOTHY A. (editor) 

1990 Bandelier Archaeological Excavation Project: Summer 1989 Excavations at Burnt Mesa 
Pueblo. Reports of Investigations No. 62. Department of Anthropology, Washington 
State University, Pullman. 

6 KOHLER, TIMOTHY A., and MATTHEW J. ROOT (editors) 

1992 Bandelier Archaeological Excavation Project: Summer 1990 Excavations at Burnt Mesa 
Pueblo and Casa del Rito. Reports of Investigations No. 64. Department of 
Anthropology, Washington State University, Pullman. 

7 WHITE, JOSEPH COURTNEY 

1992 In the Land of the Delight Makers: An Archaeological Survey in the American West. 
University of Utah Press, Salt Lake City. 



8 KOHLER, TIMOTHY A., and ANGELA R. LINSE (editors) 

1993 Papers on the Early Classic Period Prehistory of the Pajarito Plateau, New Mexico. 
Reports of Investigations 65. Department of Anthropology. Washington State 
University, Pullman. 

9. POWERS, ROBERT P., and JANET D. ORCUTT (editors) 

In The Bandelier Archeological Survey. Intermountain Cultural Resources Center 

preparation Professional Paper No. 57. Santa Fe. 



VI 



Foreword 



In 1916 Bandelier National Monument 
was established by proclamation of President 
Woodrow Wilson to protect and preserve for 
public enjoyment and education the large Pueblo 
settlements and spectacular cave dwellings of the 
southern Pajarito Plateau. At the time, the 
monument and its archaeological resources 
enjoyed considerable national prominence both 
in the public eye and within the discipline of 
archaeology, largely as a result of the pioneering 
explorations of Adolph Bandelier and the later 
excavations and preservation efforts of Edgar L. 
Hewett. Since then the monument has ceded 
much of its prominence in southwestern 
prehistory, as the focus of archaeological 
research has shifted to other regions. Although 
sporadic investigations have occurred over the 
last 75 years, the extent to which Bandelier has 
been forgotten is exemplified by the modest 
number of documented sites. In 1985 fewer 
than 500 were known in the 51 square miles of 
the monument. Knowledge of even these was 
poor at best. 

The present volume by H. Wolcott Toll 
represents the third of several National Park 
Service and Washington State University 
contributions that report the findings of the 
Bandelier Archeological Survey. Through these 
publications we hope to reestablish publicly and 
professionally the monument's important place 
in late Pueblo prehistory. The ten-year 
Bandelier Survey was begun in 1985 with the 
goal of recovering both research and cultural 
resource management data, so that the Park 
Service may better understand and interpret the 
monument's archaeology, and also better 
preserve it. 



These objectives provided the impetus 
for the present study of cavate architecture, 
undertaken in 1986 under the able direction of 
Wolky Toll. Cavates-cave rooms excavated 
into tuff cliff deposits-occur along the entire 
eastern slope of the Jemez Caldera from Tsiping 
in the north to the vicinity of Cochiti Pueblo in 
the south. Within Bandelier, cavates are 
concentrated primarily, but not exclusively, in 
Frijoles Canyon and at Tsankawi, where they 
are primarily late (post- a.d.1400 to 1600) and 
contemporaneous with the nearby, large Classic 
Period pueblos of Tyuonyi and Tsankawi. 
Outside of these settings, in the prehistoric as 
well as modern backcountry, cavates are isolated 
and smaller, ranging from single rooms to 
modest pueblos, and temporally span virtually 
the entire Pueblo occupation of the plateau (a.d. 
1 190 to 1600). Because these smaller sites were 
not known in 1986, the study necessarily 
concentrated on the larger, Frijoles and 
Tsankawi groups. 

Although cavates have drawn more 
popular attention than any other Pajaritan 
settlement type, they have received 
comparatively little scholarly study, despite the 
wealth of architectural features preserved within 
their walls. Because past work had frequently 
focused on a few rooms in a particular cavate 
group, we felt that a new, more expansive 
approach was necessary-combining recovery of 
relatively detailed information with a larger, 
multi-site sample of cavate rooms. Such a study 
would identify the range of room variability and 
function, and also document the condition of the 
rooms as a basis for future preservation. The 
sample of more than 350 cavate rooms from 4 
cavate groups in Frijoles Canyon and 1 cavate 



Vll 



group at Tsankawi provides a strong foundation 
for archaeological inference and preservation 
planning. 

Because the focus of the investigation is 
the architecture of cavate rooms, the study 
differs in important ways from the 1987-1991 
inventory survey. The inventory goal of 
surveying 40 percent of the park required 
investigation at the broader scales of component 
and site, an approach that minimized our ability 
to examine smaller units, such as individual 
rooms and their features. In this respect the two 
studies are at once different and complementary. 
Four of the five cavate groups recorded in 1986 
and reported here were later inventoried by the 
survey. At these groups the availability of both 
small- and large-scale architectural data provides 
a level of architectural documentation rivaled 
only at inventory sites later excavated by Tim 
Kohler of Washington State University. Plane 
table plan and profile views of three of the 
Frijoles cavate study groups (Groups F, I, and 
M) prepared during the inventory have been 
included in the present volume. Original 
sketches produced during 1986 for the remaining 
two sites (Group A and Tsankawi) have been 



retained, since the first of these sites was not 
included within the inventory sample areas, 
while the second was recorded, but not 
instrument-mapped . 

Rapid, systematic, and comprehensive 
recording of several hundred rooms containing 
not only floor and wall features but also roof 
features is a professional challenge most of us 
have never faced, but one that Toll has met with 
ingenuity. One particular problem was how to 
document hundreds of often dark interior 
surfaces in a manner that would establish a 
condition baseline that could be used by future 
investigators to measure deterioration. 
Videotaping each room, with a running audio 
commentary, was the solution. Many field 
projects should evaluate this technique as a 
backup to notes, maps, and still photography. 
Metal-based videotape, now widely available, 
provides long, nearly archival tape life, and tape 
digitization ensures a virtually permanent record. 

Robert P. Powers, Director 

Bandelier Archeological Survey 

June 1994 



vm 



Acknowledgments 



The number of people participating in 
the cavate recording project was pleasantly 
small, so that each person's contribution forms 
a substantial part of the result. The fieldwork 
was carried out full time by Barbara Mills, 
Bruce Panowski, and myself. Working with 
Barbara and Bruce was a genuine pleasure. 
Barbara was a steady source of thoughtful 
input, hard work, and good spirits; Bruce 
accepted his consignment to the field and 
commuting with his usual good cheer. 
Especially warm and heartfelt thanks are due to 
the four volunteers who all traveled great 
distances in various ways to contribute their 
skills and time. Bill and June Crowder, Betsy 
Fuller, and Liz Bayer all helped us accomplish 
a great deal more than would have otherwise 
been possible. The park staff was uniformly 
pleasant and helpful, making Bandelier an even 
nicer place to work. Ed Greene's rapelling into 
a few hard-to-reach cavates kept us all in 
suspense in more than one way. Peter McKenna 
and Bob Powers conducted surface inventories 
of the cavates to attempt to date the cavate 
groups studied. Peter provided the section on 
ceramics and dating and continued to provide 
advice on chronology and ceramics during 
preparation of this report. The Crowders 
quickly put all their photos and records in good 
order and June prepared the summary of her 
detailed rock art study. 

Bruce Panowski continued to spend a 
good deal of time on the project after we left the 
field. He coordinated the data-entry phase, 
helped to locate and fix problems, and generated 
the preliminary outputs. He and Tony 
Tagliaferro also helped bring the data back to 



the NPS system after its sojourn at UNM. Mary 
Padilla, Suzette Lopez, Sophia Ulibarri, and 
Betsy Fuller all struggled nobly with entering the 
forms into the computer (especially those left- 
handed ones). Several members of the staff of 
the University of New Mexico Computing 
Center were very helpful. Without Mike Prine's 
help, there is some question as to whether the 
data could have been transferred to the UNM 
system at all. Dusty Teaf and Sandy Robinson 
cheerfully and energetically assisted with tape 
and disk management problems and SAS 
questions many times. 

Robert Preucel made extensive and 
helpful comments on the draft and shared data 
and information from his involvement with the 
Pajarito Archaeological Research Project of the 
University of California at Los Angeles. Dr. 
Lys Ann Shore provided a careful, in-depth 
technical edit of the whole volume. Jerry 
Livingston drafted the final figures, and, assisted 
by Ernesto Martinez, helped with interim maps. 
Anne Goldberg did extensive typing and 
formatting of tables for the final version, and 
Sarah Chavez put the whole thing into camera- 
ready format. Kathleen Havill did the indexing. 

Bob Powers was involved with the 
project throughout. As director of the Bandelier 
Survey, he initiated and administered it. He 
oversaw the contracts for write-up and 
completion, read drafts and made suggestions. 
In spite of ever-increasing duties at the Park 
Service, Bob continued to put in a great deal of 
time and thought during the editing and 
production phases of the report, exhaustively 



IX 



checking the text, and carefully redrafting 
figures. I especially appreciate his having 
arranged for follow-up contracts to cover my 
time during final edits. 

The Office of Archaeological Studies, 
Museum of New Mexico allowed me to take 
time off to work on this report on numerous 
occasions. I was gone or grouchy at several 



points along its way (they were presumably glad 
when the two were simultaneous) Mollie Toll 
and our sons Nick and Spencer deserve thanks 
for their patience, understanding, and support 
during all phases of my involvement with the 
project and its write-up. Spencer was not yet 
born when the fieldwork was done for this 
report; he can now ride a bicycle. 



CONTENTS 



FOREWORD BY R. P. POWERS vii 

ACKNOWLEDGMENTS ix 

LIST OF FIGURES xiii 

LIST OF TABLES xv 

1. INTRODUCTION TO CAVATES AND THEIR STUDY 1 

What Is a Cavate? 1 

Why Study Cavates? 2 

Past Work in Cavate Features 2 

The Present Project 14 

2. CONTEXT, DESCRIPTIONS, AND CHRONOLOGY 15 

The Setting 15 

The Sites 17 

Cavate Chronology 61 

Incorporation of Ceramic Dates and Further Dating Potential 61 

Analysis of Surface Ceramics from the Study Areas by Peter J. McKenna 64 

3. RECORDING PROCEDURES, GROUP ATTRIBUTES, AND CAVATE 

CONDITION 77 

Recording Procedures 77 

Data Manipulation 90 

Group Attributes 93 

Cavate Condition 99 

4. CAVATE AND NONCAVATE FEATURES: DEFINITIONS, DISTRIBUTIONS, AND 
DIMENSIONS 107 

Structural Features 108 

Floor Features 138 

Wall Features 150 

Summary of Detailed Rock Art Study by June Crowder 197 

5. PRELIMINARY FUNCTIONAL ANALYSIS OF CAVATE CHAMBERS 201 

Feature Co-occurrence 201 

Plastering and Smoking 206 

Cluster Analysis 206 



XI 



6. INTERPRETATION AND CONCLUSION 213 

Cavate Use and Relationship to Large Surface Pueblos 216 

Conclusion 217 

APPENDICES 219 

1. Forms, Coding, and Materials Collected 221 

2. Data Sets, Volume Calculation, Output Listing, and Photographic Data 237 

3. Base Information, Threatened Cavates, and Room Stability 245 

4. Detailed Listing of Rock Art and Historical Correlation with Chapman 

by June Crowder 267 

5. Chamber Cluster Membership 275 

REFERENCES 285 

INDEX 293 



XII 



LIST OF FIGURES 



1.1. Map showing areas of the Pajarito Plateau in which cavates have been 

studied, and the extent of the Bandelier Tuff 3 

2.1. Map of Frijoles Canyon showing the location of the groups recorded in 1986 . . 18 

2.2. Map of the Tsankawi section, including the vicinity of the group 

of cavates recorded in 1986 19 

2.3. Cross-canyon view of Group I 21 

2.4. Cross-canyon view of upper Group M, the portion recorded in 1986 21 

2.5. Room plan view for the Group A sample 23 

2.6. Elevation sketch for the Group A sample 25 

2.7. Comparison photos for A-13, 1939-1986 30 

2.8. Comparison photos for A-60, 1939-1987 31 

2.9. Masonry front wall of A-10, 1986 32 

2.10. Room plan view for the Group F sample 33 

2.11. Elevation sketch for the Group F sample 35 

2.12. Comparison photos for F-31, 1939-1986 38 

2.13. Comparison photos for upper Group F, Room 12, 1939-1986 39 

2.14. Room plan view for Group I 41 

2.15. Elevation sketch for Group I 43 

2.16. Comparison photos for 1-22, 1939-1986 46 

2.17. Room plan view for the Group M sample 47 

2.18. Elevation sketch for the Group M sample 49 

2.19. Room plan view for the Tsankawi sample . 55 

2.20. Elevation sketch for the Tsankawi sample 57 

2.21. Comparison photos for TS-53, 1939-1986 60 

3.1. Schematic drawing of a cavate showing examples of several of the79 

features recorded 79 

3.2. Example of digging stick marks in A-10 82 

3.3. Evidence of chamber expansion in M-10 83 

4.1. Histogram showing volume distribution for all chambers at Frijoles 

and Tsankawi 117 

4.2. Histogram showing volumes for all chambers with assigned functions 118 

4.3a. Histogram comparing volumes for habitations and "kivas" at Frijoles 

and Tsankawi 119 

xiii 



4.3b. Histogram comparing volumes for storage rooms at Frijoles and Tsankawi .... 120 

4.4. Floor depressions/pot rests and large wall niche in A-47 142 

4.5. Floor pit complex in TS-55 144 

4.6. Metate rest and floor ridge in M-40 145 

4.7. Grinding complex in M-60 146 

4.8. Row of loom anchors in M-59 148 

4.9. Wooden loom anchor loop in TS-59 148 

4.10. Adobe collar and wall niche in F-37 151 

4.11. Deflector in upper Group F, Room 15 152 

4.12. Large floor-level niches 153 

4.13. Large floor-level niche volumes, showing Frijoles and Tsankawi cases 156 

4.14. Plot of height above floor and volume of wall niches 160 

4.15. Example of a slot in M-59 162 

4.16. Three-dimensional plot of viga hole height, depth, and diameter 166 

4.17. Three-dimensional plot of "latilla" hole height, depth, and diameter 168 

4.18a. Three-dimensional plot of beam seat height, depth, and diameter 169 

4.18b. Three-dimensional plot of viga hole and beam seat height, depth, 

and diameter 172 

4.19a. Diameters of cavate indeterminate holes, showing Frijoles 

and Tsankawi cases 175 

4.19b. Depths of cavate indeterminate holes, showing Frijoles and Tsankawi cases . . . 176 

4.20a. Three-dimensional plot of height, depth, and diameter of indeterminate holes . . 177 

4.20b. Three-dimensional plot of indeterminate holes with diameters and depths 

of 10 cm or less 178 

4.21. Plot showing overlap of viga, latilla, and indeterminate distributions 181 

4.22. Narrow wall incisions in TS-24 189 

4.23. Cliff niches on extramural cliff at Tsankawi 191 

4.24. Rock art, viga holes, and plaster dado in TS-59 192 

4.25. Large bird figures beside door to TS-40 193 

4.26. Bird figure in F-23 193 

4.27. Awanyus in M-13 194 

4.28. Plastered-over masks in M-60 194 



xiv 



LIST OF TABLES 



1.1. Group M ceramic types reported by Turney (1948) with counts from 

cavate surface material analysis 10 

2.1. Chronometric dates and ceramic associations from Bandelier 63 

2.2. Date ranges used by various analysts for Rio Grande ceramic types 

2.3. Ceramic samples for Bandelier cavates 68 

2.4. Ceramic form and ware data and significance tests 73 

3.1. Distribution of noncavate feature types 85 

3.2. Time spent and forms completed by cavate group, 1986 89 

3.3. Room type and mode of recording by site 94 

3.4. Evidence for construction 95 

3.5. Estimate of excavated versus natural space (cavates only) 96 

3.6. Masonry presence and type 97 

3.7. Room level, cavates and noncavates combined 98 

3.8. Fill type, cavates only 98 

3.9. Fill depth, cavates only 98 

3.10. Tuff characteristics, cavates only 99 

3.11. Nonhuman use of cavates 100 

3.12. Overall stability of cavates and noncavates 100 

3.13. Human damage by group and chamber location 102 

3.14. Natural damage by group and chamber location 104 

4.1. Overall occurrence of cavates, noncavates, and features by group 108 

4.2. Occurrence of all individually recorded features in cavates and noncavates, 

by cavate group (number and percentage) 109 

4.3. Chamber occurrence and dimensions 113 

4.4. Exterior door occurrence and dimensions 121 

4.5. Exterior opening occurrence and dimensions 123 

4.6. Occurrence of doors and openings by group and cavate type 124 

4.7. Interior door occurrence and dimensions 125 

4.8. Natural wall occurrence and dimensions 127 

4.9. Plaster coats by group, wall, and function 129 

4.10. Mean plaster height and number of coats by group and wall 131 

4.11. Plaster color by group 132 

4.12. Masonry wall occurrence and dimensions 133 

4.13. Floor occurrence and dimensions 135 

xv 



4.14. Floor plaster coats by group and function 136 

4.15. Ceiling occurrence and smoking by group and function 137 

4.16. Firepit occurrence and dimensions 139 

4.17. Floor pit occurrence and dimensions 141 

4.18. Floor depression occurrence and dimensions 143 

4.19. Metate rest occurrence by location and group 147 

4.20. Co-occurrence of possible mealing complex features 147 

4.21. Loom anchor occurrence and dimensions 150 

4.22. Large floor-level niche occurrence and dimensions 154 

4.23. Wall niche occurrence and dimensions 158 

4.24. Co-occurrence of floor-level and wall niches by group 161 

4.25. Slot occurrence and dimensions 163 

4.26. Viga hole occurrence and dimensions 164 

4.27. "Latilla" hole occurrence and dimensions 167 

4.28. Beam seat occurrence and dimensions 170 

4.29. Indeterminate hole occurrence and dimensions 173 

4.30. Results of cluster analysis on round wall holes 180 

4.31. Discriminant analysis classification of feature types 180 

4.32. Loom support occurrence and dimensions 183 

4.33. Smokehole occurrence and dimensions 184 

4.34. Vent occurrence and dimensions 186 

4.35. Groove occurrence by shape and group 186 

4.36. Wall depression occurrence and dimensions 187 

4.37. Rock art occurrence and chamber location 195 

4.38. Cavates containing rock art by group and motif 198 

4.39. Rock art nomenclature 199 

5.1. Summary of chamber attribute occurrence 202 

5.2. Co-occurrence of feature categories in chambers 203 

5.3. Spearman rank-order correlations of feature category co-occurrence 204 

5.4. Plaster coat- feature number co-occurrence 207 

5.5. Smoking of plaster coats 208 

5.6. Means, membership, and correspondence to assigned function for chamber 

cluster analysis 209 

6.1. Frequencies of features by assigned function 214 



xvi 



Introduction to Cavates and Their Study 



The true character of the so-called "Cavate lodge" has not been fully understood. 

E. L. Hewett, 1909 



The Pajarito Plateau of New Mexico was 
formed largely by a series of gigantic volcanic 
ash flows. The ash consolidated into a soft rock 
called tuff, which was gradually dissected into 
many deep canyons by runoff from the Jemez 
Mountains to the Rio Grande. In the sheer cliffs 
formed by the tuff are literally thousands of 
"cavates," chambers hollowed out of the tuff by 
the prehistoric Pueblo people who flourished 
there in the twelfth through sixteenth centuries. 
Bandelier National Monument contains several 
canyons with abundant cavates. This study is 
the result of a pilot project to investigate this 
one type of archaeological feature in detail, as 
part of a larger project to inventory the park's 
archaeological resources. 

What Is a Cavate? 

It seems likely that archaeology provided 
the word cavate to the language, at least as a 
noun. Webster's definition of the term is "cut in 
soft rock: EXCAVATED < -cliff dwelling > ." 
Recent dictionaries, however, show the word 
only as a verb meaning "to hollow out" and list 
its use as "rare"; given this apparent trajectory, 
it seems archaeologists had better use the term 
or lose it. It is not known when the term first 
came into archaeological use, though Mindeleff 
(1896:217) provides a useful discussion of it as 
early as 1896: 



Cavate lodges comprise a type of 
structures closely related to cliff houses 
and cave dwellings. The term is a 
comparatively new one, and the 
structures themselves are not widely 
known. They differ from the cliff 
houses and cave dwellings principally in 
the fact that the rooms are hollowed out 
of cliffs and hills by human agency, 
being cut out of soft rock, while the 
former habitations are simple, ordinary 
structures built for various reasons 
within a cove or on a bench in the cliffs 
or within a cave. 

The term seems to have had considerable 
currency around the turn of the twentieth 
century (e.g., Powell 1886, 1891; Mindeleff 
1896; White 1904; Bierbower 1905; Beam 
1906). 

In this study, cavate features are defined 
as cavities in the canyon wall that are primarily 
the result of excavation of the rock. Both 
Mindeleff and Hewett recognized that there is 
variability in these features, how they are 
incorporated into structures, and how they relate 
to other sites lacking cavates. Hewett 
(1909a:438), however, contended that "this term 
is one that should be rejected from the 
nomenclature of Southwestern archeology." 
Siding with Mindeleff (1896) and Fewkes 



CAVATE STRUCTURES 



(1913:193 Note 1), our position is that cavates 
are sufficiently different from other 
archaeological remains to be a useful separate 
category (see also Hall 1992:23-24). This is 
especially so in an area such as the Pajarito 
Plateau where these features abound. 

Although it is not possible at present to 
draw the absolute limits of cavate distribution in 
northern New Mexico, there is little doubt that 
the line would follow the boundaries of the 
Bandelier Tuff, which is more or less 
coextensive with the Pajarito Plateau (Figure 
1.1; Bailey et al. 1969; Ross et al. 1961; 
Mathien et al. 1993). On the north, cavates are 
present at the Tsiping Ruin at the north end of 
the Jemez Mountains near Cerro Pedernal 
(Dougherty 1980:17; Stuart and Gauthier 
1981:104). It is unlikely that any cavates exist 
east of the Rio Grande. They are present as far 
west as the Jemez River and appear to extend 
south of 3andelier National Monument (Fliedner 
1975; R. Preucel, personal communication, 
1988), at least to Peralta Canyon northwest of 
Cochiti Pueblo. 

Why Study Cavates? 

The study of cavates serves two main 
purposes: adding to archaeological knowledge of 
these features and assessing their current 
condition as a means of caring for them. These 
concerns depend on the basic archaeological goal 
of recording these features in a systematic way 
in order to assess their variability. Knowledge 
of the variability of cavates will help 
archaeologists and interpreters understand the 
prehistoric functions of these features and may 
enable them to date cavates more precisely and 
determine whether the features changed through 
time. Cavate structures are unusual in the 
prehistoric Pueblo record, and they are better 
suited to recording without excavation than most 
architectural features. Cavates are much like 
dry caves, so that many fragile features- 
including plastering and even organic materials- 



are remarkably well preserved in cavates. 
Cavates also preserve information that is seldom 
available in excavated sites, since ceilings and 
full walls are present. This permits better 
estimation of room volume than is usually 
possible. Cavates have a great deal of 
archaeological information available for 
collection with little disruption of deposits and 
relatively little effort. In spite of the attention 
cavates have received, there has been 
surprisingly little formal recording of them. 
They are so numerous on the Pajarito Plateau 
that a thorough yet efficient means of recording 
them is essential to a well-rounded survey of 
Bandelier and the Pajarito Plateau. 

On the management level, recording the 
cavates provides a basis for monitoring them for 
deterioration and for devising and assessing 
countermeasures. Cavates attract considerable 
attention from both visitors and vandals, 
subjecting them to both casual and intentional 
attrition. They are also probably more subject 
to several types of natural degradation than are 
other archaeological sites: more features are 
either exposed or very shallowly buried, the tuff 
is very friable, and there is possibility of 
collapses of canyon wall segments. 

Past Work in Cavate Features 

Cavates in Other Areas 

Ideally, the cavates found in Frijoles 
Canyon should be placed in a broad 
anthropological perspective, but that endeavor is 
beyond the scope of this study. Rooms 
excavated out of the rock exist elsewhere in the 
Southwest and around the world. Studies of 
cavates in other places may contain information 
on labor, ventilation, storage, insulation, and 
maintenance relevant to those of the Pajarito 
Plateau. Kempe (1988) provides a more global 
view of cave dwellings, including substantial 
material on artificial caves, emphasizing Old 
World sites. Although he includes a chapter on 



INTRODUCTION 




San Juan 



Santa Clara 



San lldefonso 



it 



Tsankawi Pueblo 
FIG. 2-2 
Tshirege 

Mortendad Canyon 
FIG. 2-1 



-+■ Peaks 

^^ Limits of Bandolier Tuff 

A Cavate Groups in This Study 

A Cavate Groups Mentioned 

Modern Pueblos 
Towa S Keres KS Tewa 



Santo Domingo 



o 5 

M l l I l 



25 



30 kilometers 

_l 



Figure 1.1. Map of the Pajarito Plateau and environs, showing the locations of study areas (areas 
of figures 2.1 and 2.2 are indicated by dotted lines) and other recorded cavates, the 
distribution of the Bandelier Tuff, the Valle Grande Caldera, major streams, and modern 
pueblos. The elevation of the Rio Grande at the mouth of the Rito de los Frijoles is 
1646 m. (Compiled from Bailey et al. [1969]; Dane and Bachman [1965]; Dougherty 
[1980]; Hyland [1986]; Ross et al. [1961]; USDA Forest Service Santa Fe National Forest 
Map). 



CAVATE STRUCTURES 



the Four Comers area of the American 
Southwest, he does not mention the cavates of 
the Pajarito Plateau, in spite of their similarity to 
excavated structures he discusses in other parts 
of the world. 

Some striking parallels to the Pajarito 
Plateau can be found in Cappadocia, an ancient 
province in what is now central Turkey. The 
parallels are primarily geological rather than 
archaeological: Cappadocia has large expanses 
of tent rocks formed in a thick layer of volcanic 
tuff (Blair 1970; Heiken 1979; Fewkes 1910; 
Kempe 1988; Riboud 1958; Severy 1983:737). 
Cut into this tuff are thousands of rooms, 
ranging from small monastic cells to elaborate 
churches. From a distance many of the smaller 
rooms look very much like Bandelier cavates, 
but on the whole they are probably larger and 
more elaborate and ornate (e.g., Heiken 
1979:568-569). Some of these sites are truly 
grand in scale; for example, "Vardzia numbers 
at least five hundred rooms and apartments, 
including chapels, banqueting halls, wine cellars, 
stables, all connected with a labyrinth of stairs 
and passages. It is made up of a number of 
storeys [sic], being cut out of a high vertical 
cliff face" (Lang 1966:125, also Plate 55). 

Cosmos Mindeleff (1896) lists four areas 
of the Southwest where cavates are found: the 
Rio Grande, the lower San Juan River drainage 
including the lower Mancos River, the Flagstaff 
area, and the Verde Valley. Both the San Juan 
and Verde examples are excavated out of soft 
sedimentary deposits, while those in the 
Flagstaff area are dug into "cinders." Mindeleff 
(1896:217-235) notes that there are thousands of 
cavates in the Verde Valley, usually in small 
groups, but sometimes in groups of several 
hundred rooms. He devoted considerable care 
to recording some of the cavates in the Verde, 
including several individual structure plans, 
measurements, a map of a very large cluster, 
and photographs. Fewkes (1913:188-193) 
recorded other cavate rooms further up the 
Verde drainage, and discussed their relationship 



to nearby masonry structures. Mindeleff reports 
rooms as large as 3.7 x 6.7 m and series of 
rooms extending up to 14 m back into the bluff, 
far deeper than any artificial caves we 
encountered on the Pajarito Plateau. 

Recording of cavates in the Verde 
Region, desultory since the time of Mindeleff 
and Fewkes (see Hall 1992:50-66), took a major 
step toward rigor with Susan Hall's (1992) 
recording and analysis. Hall's thesis focusses on 
the Mindeleff Cavate Site, providing plans and 
elevations for a large number of cavate suites 
and data on feature types and room sizes. In all, 
she collected data from 343 rooms in 89 cavate 
suites (Hall 1992:69). Hall also provides 
theoretical direction for the architectural study of 
cavates. Principle differences between the 
Verde cavates (at least at the Mindeleff Cavate 
Site) and those on the Pajarito Plateau seem to 
be that there is less use of exterior masonry on 
the Verde, and that the Verde examples are 
larger and have much more complex cavate 
plans, with many chambers linked together from 
a single exterior opening. Single chambers with 
exterior openings are the norm in the cavates we 
studied, with three chambers the maximum and 
the exception; in Hall's (1992:104) sample, most 
suites contained 2 to 5 rooms, and she recorded 
two cases with 10. Some of these rooms are 
small enough that we might have called them 
large niches, but the differences are marked. 

Quoting the peripatetic Major Powell 
(1891), Mindeleff (1896:223-224) gives a brief 
description of two groups of structures east of 
San Francisco peak near Flagstaff. One of these 
includes about 150 rooms dug into a cinder 
dome, the top of which was walled and levelled 
to form a plaza. The cavate rooms were 3 to 4 
m in diameter and 2 to 3 m high, arranged as a 
larger chamber central to smaller ones. The 
second group was built around the crater of a 
larger cone. Here a combination of free- 
standing masonry, utilized natural caves, and 
excavated caves formed a village which also had 
a formal plaza. Powell postulated that the 



INTRODUCTION 



builders of these villages were ancestors of the 
Havasupai. 

Fewkes (1904:35-39; Whittaker n.d.) 
describes three different techniques of cavate 
construction in the Flagstaff area, providing 
some photos, a plan, and more detail than 
Mindeleff. At Turkey Tank Caves there are 
alternating layers of hard lava and breccia and 
sections of breccia have been removed, with the 
resulting cavities partitioned with walls. At the 
New Caves site, a number of cavate rooms were 
dug into the steep walls of an extinct crater, with 
masonry rooms in front. At Old Caves, there is 
an extensive single-story masonry pueblo many 
rooms of which have a subterranean component. 
Some of these "cellars" contain several chambers 
(see Fewkes 1904:37). Colton (1932:25; 
1946:37-39) also recorded and mapped Old 
Caves, noting that "the curious underground 
chambers in nearly every room hollowed out of 
the cinders make it unique in pueblo 
architecture" (Colton 1932:25). He dates these 
structures to Late Pueblo III, a.d. 1250-1300. 

Cavates are uncommon in the San Juan 
area. Prudden (1903:252-253) observes that 
"these examples [in the Mancos River drainage] 
are so small or so weathered away that one who 
should be tempted to make the long journey to 
the San Juan or lower Mancos for the sake of a 
study of cavate lodges would risk 
disappointment, especially in view of the more 
extensive, varied, and typical groups in the 
Verde valley ... or those in the valley of the 
Rio Grande upon the eastern slope of the 
Valles." Prudden includes a photograph of a 
"weathered remnant of one of the cavate 
dwellings, showing the soft shale in which the 
shallow shelters were dug" (Prudden 1903: Plate 
XXX). 



Pajarito Plateau 

For a variety of reasons, people have 
been in and out of cavate structures on the 



Pajarito Plateau ever since their abandonment. 
Hendron (1943) found evidence of reuse in the 
seventeenth century (ca. Pueblo Revolt?) and 
later. Both the Tewa and the Keres have clearly 
used the area for centuries, and there can be 
little doubt that the caves were visited 
periodically after the large sites were abandoned 
and before there was a significant non-Indian 
presence in the area. 

Extensively recorded visits probably 
began with Adolph Bandelier, who first visited 
Frijoles Canyon in 1880, making collections and 
observations. On his second visit in December, 
he slept in cavates in Alamo Canyon and near 
Tyuonyi. He found them quite comfortable in 
spite of some cold weather (see Lange and Riley 
1966, especially 225-228). He appears to have 
selected large chambers in which to stay: he 
gave two height measurements of more than 2 
m. He referred to the cavate where he stayed in 
Frijoles as the Room of the Cacique. At that 
time the rooms were well preserved and the 
ruins in general "very rich in fine fragments of 
pottery and manos"; the pottery was 
"prevailingly glazed." He visited many cavates 
and measured several of them. Many of his 
observations still hold, though artifacts and 
masonry are now considerably less abundant. 

The rooms are remarkably well 
preserved in most cases, and much 
stonework used. The goats have filled 
them with their dung. . . The floor is 
perfect in most cases, also the yellow 
plastering. The ceiling is generally 
smokey [sic] and sooty. . . There are, 
lower down, several of these large 
circular rooms like our present quarters. 
Were they estufas? The Indians say not; 
they are all houses, and the estufas were 
those below in the valleys. . . Every 
room has its fireplace, except such as 
were evidently used as storehouses. The 
ruins are in groups, and the deep 
recesses and reentering angles of the 
cliffs are avoided. 



CAVATE STRUCTURES 



... In general the rooms of the 
eastern half [presumably of the cavate 
group in which they were staying] are 
larger than those of the western section, 
there are even a number of very large 
ones. They are all plastered yellow, and 
smoky above. . . 

Many of the rooms contain 
carved walls, but while the carvings 
may have been made by Indians, they 
are certainly posterior to abandonment 
of the caves, as they are carved in the 
plastering. (Lange and Riley 1966:226- 
228) 

Bandelier returned in 1885, checking his 
descriptions for Die Koshare (later translated and 
expanded into The Delight Makers). At that 
time there were three ranchos in the valley. 
One of his companions, named Pacifico, "found 
a black olla, entire, in one of the caves; also two 
stone axes" (Lange et al. 1975:76). They 
proceeded further north, noting large artificial 
caves in "Canada Ancha" and many caves at 
"Tzirege." In 1886 Bandelier visited Puye, 
where he further showed his interest in caves: 

On the whole they are an exact 
repetition of the Rito, only their 
situation is different. They are plastered 
with yellow clay, and there is a smoke 
escape cut out above the doorways. 
Floors are everywhere black and about 
two inches thick. Many holes for 
beams. On the average, they are only 
one story, but I saw two and three 
stories also. There are also beam holes 
indicating porches in front of the rock. 
The caves are singularly distributed, and 
they are high, over all timber and 
plainly visible at a great distance. It is 
a good position for defense and watch. 
(Lange et al. 1975:160) 

Bandelier detailed his vision of life and social 
organization in the cavates in The Delight 
Makers (1971), published in 1890. 



At about the same time that Bandelier 
was pursuing his investigations, members of the 
Bureau of American Ethnology visited cavates 
on two occasions and briefly stated their findings 
in annual reports of that organization. James 
Stevenson spent a month in Frijoles Canyon in 
1882. 

In many of the caves which were 
examined a flooring of fine red clay, 
very neatly and smoothly spread in 
several thin layers, is still seen, as also 
a plastering of red or yellow clay upon 
the walls. In some of them the lower 
part of the wall is of one color and the 
upper part and ceiling of another, the 
two colors being separated by a broad 
line of dark brown or black which runs 
around the cave about two feet from the 
floor. In the walls were found small 
niches. 

Beneath some of these caves, 
which were situated higher in the face of 
the cliff, were evidences of the former 
existence of annexed exterior chambers 
below. The cliff walls beneath these 
apertures had evidently been hollowed 
out to form the rear wall of the annexed 
chamber, and were nicely plastered with 
red and yellow clay. Rows of small 
round holes were seen which, it was 
thought, had been used as rests for the 
rafters, while large quantities of roughly 
squared stones used in building lay 
scattered about the base of the cliff. In 
some cases there appeared to be two and 
even three tiers of houses constructed in 
this manner. (Powell 1886:xxxvi- 
xxxvii) 

J. W. Powell, the director of the bureau, 
visited the cavates (which he called by that term) 
in 1886. Powell (1891:xxi-xxiv) devoted several 
pages to discussing cavates. He considered them 
to be dwellings reached either by ladders or 
artificial terraces. He noted the presence of 
firepits, niches, and abundant potsherds. He 



INTRODUCTION 



made some chronological interpretations that 
differ from those now held: 

On more careful survey it was found 
that many chambers had been used as 
stables for asses, goats, and sheep. 
Sometimes they had been filled a few 
inches, or even two or three feet, with 
the excrement of these animals. . . 
Altogether it is very evident that the 
cliff houses have been used in 
comparatively modern times; at any rate 
since the people owned asses, goats, and 
sheep. The rock is of such a friable 
nature that it will not stand atmospheric 
degradation very long, and there is 
abundant evidence of this character 
testifying to the recent occupancy of 
these cavate dwellings. . . Every mesa 
had at least one ancient pueblo upon it, 
evidently far more ancient than the 
cavate dwellings found in the face of the 
cliffs. It is, then, very plain that the 
cavate dwellings are not of great age; 
that they have been occupied since the 
advent of the white man, and that on the 
summit of the cliffs there are ruins of 
more ancient pueblos. . . It was at once 
noticed that the potsherds of these cliff 
dwellings are, both in shape and 
material, like those now made by the 
Santa Clara Indians. . . While 
encamped in the valley below, the party 
met a Santa Clara Indian and engaged 
him in conversation. From him the 
history of the cliff dwellings was soon 
obtained. His statement was that 
originally his people lived in six 
pueblos, built of cut stone, upon the 
summit of the mesas; that there came a 
time when they were at war with the 
Apaches and Navajos, when they 
abandoned their stone pueblos above and 
for greater protection excavated the 
chambers in the cliffs below; that when 
this war ended part of them returned to 
the pueblos above, which were rebuilt; 



that there afterward came another war, 
with the Comanche Indians, and they 
once more resorted to cliff dwellings. 
At the close of this war they built a 
pueblo in the valley of the Rio Grande, 
but at the time of the invasion of the 
Spaniards, their people refused to be 
baptized, and a Spanish army was sent 
against them, when they abandoned the 
valley below and once more inhabited 
the cliff dwellings above. Here they 
lived many years, until at last a wise 
and good priest brought them peace, and 
persuaded them to build the pueblo 
which they now occupy-the village of 
Santa Clara. . . It is therefore evident 
that the cavate dwellings of the Santa 
Clara region belong to a people still 
extant; that they are not of great 
antiquity, and do not give evidence of a 
prehistoric and now extinct race. 
(Powell 1891:xxiii-xxiv) 

In the early twentieth century a visit to 
the Pajarito Plateau became a vacation 
adventure. Susan Bierbower (1905) reported her 
visit to Puye. Her Santa Clara guide was 
apparently an astute, early cultural-resource 
conservationist, and Bierbower herself something 
less than a cultural relativist: 

The next morning . . . armed 
with a good staff and my kodak [sic], I 
again ascended to the dwellings. We 
had provided ourselves with a pick and 
shovel for excavating, and you can 
judge of our disgust when we 
ascertained that Juan had left them in 
Santa Clara. It is my firm belief that 
this was done with malice aforethought, 
as we learned that the Indians are very 
superstitious and unwilling to disturb 
these places. A small trowel was all we 
had. (Bierbower 1905:232) 

Beam (1906) produced a more widely 
circulated report on the Pajarito, including 



8 CAVATE STRUCTURES 



several photographs of cavates at Puye (despite 
his title, "The Prehistoric Ruin of Tsankawi"). 
Both Bierbower and Beam concluded that the use 
of the cavates and pueblos in the region had 
ceased due to cataclysms, respectively 
earthquakes and "fierce and implacable enemies" 
or perhaps earthquake or eruption (Beam 
1906:813). 

The work done by Hewett in 1908 and 
1909 in Frijoles Canyon included "clearing" 
cavate rooms, for some of which he gives mea- 
surements and locations (Hewett 1909a,b). 
Hewett also defined 13 groups of cavate/talus 
sites (A-M) along the north side of the canyon. 
These designations have been used by most later 
workers in Frijoles Canyon, and they are shown 
on the remarkable and still useful map and ele- 
vation prepared by Kenneth Chapman and 
published both by Hewett (1909a, 1938) and by 
Hendron (1940). 

Hewett discussed cavates in several 
places. In the first of two 1909 papers in 
American Anthropologist he presented some ex- 
ceptional panoramic photos of large segments of 
the north wall of Frijoles, showing three of the 
groups in which our crew later worked (A,F, 
and I). He also reproduced Chapman's recon- 
struction of Long House and the interior of a 
cavate in use. An interior photo illustrated 
many features of cavate rooms, such as digging 
stick marks, differential plastering leaving a 
band, a large floor-level niche, and a possible 
upper loom support. In the second article 
(Hewett 1909b) Hewett gave plans and mea- 
surements for some cavates in the eastern part of 
Group E (Sun House), but not for the others that 
he dug in the western part of Group E (Snake 
Village). Aside from a burial in the Snake 
Village portion, he made no mention of the con- 
tents or features of these rooms. He gave more 
attention to two "kivas," larger, heavily smoked 
rooms with loom anchors in the floor. The 
Snake Village received its name from an 
Awanyu painted on the plaster of the associated 
"kiva." Hewett also partially excavated an even 



larger cave kiva in the area but did not give its 
exact location. Hewett's later publication (1938) 
contains much of the material in the 1909 art- 
icles; in addition, it presents a sort of develop- 
mental sequence based on his estimation of 
quality of workmanship, in which a chrono- 
logical element is implied but not specified. 

The following discussion of burial 
placement suggests what Hewett may have found 
but did not report in detail: 

Crypt or cave burial was here secon- 
dary. Mortuary crypts were posterior 
chambers to pueblo-like cliff dwellings. 
They were receptacles for great quan- 
tities of disjointed bones, the rooms 
being filled with these unrelated remains 
to a depth of several feet. No utensils 
accompanied them. I consider these 
crypts to have been depositories for 
bones removed from, or washed out of, 
the cemeteries above. In individual cave 
burial as practiced in this region the 
dead are found in embryonic position 
and usually wrapped in feather robes or 
matting of yucca fiber. (Hewett 
1938:134-135) 

Chapman (1916, 1938) made a survey of 
the "cave art" of the "region of the Rito de los 
Frijoles;" he gives a breakdown of 106 
prehistoric "works" by subject. He seems to 
have kept more extensive records as well, and 
he mentions an illustrated presentation, 
published at least in part as an appendix to 
Hewett's (1938) book on the Pajarito (see also 
Chapter 4 and appendix 4 of the present study). 
Chapman was especially intrigued by naturalistic 
figures and scenes scratched into plaster. While 
Bandelier attributed this style of rock art to 
postoccupational visits, Chapman clearly thought 
that at least some examples were done by 
residents of the cavates. 

Hewett and Morley also worked at Puye, 
where a large number of cavate rooms are 



INTRODUCTION 



located below the mesa-top room blocks (Hewett 
1908; Morley 1910). Stewart Peckham, 
formerly of the Laboratory of Anthropology, has 
devoted considerable effort to assembling 
Morley's notes from the Puye work, but in all 
his searches he has found no notes relating to the 
cavates (S. Peckham, personal communication, 
1987). M. R. Harrington of the Southwest 
Museum worked in cavates in 1926; Peckham 
has found a map of several cavate rooms on 
three levels, with masonry rooms in front of the 
first level. Once again, however, extensive 
research and inquiry have revealed no further 
notes. Harrington's map shows floor ridges 
(which he calls sleeping ridges), "cook tables," 
firepits, partitions, vents, and three burials, one 
in a small, sealed back chamber. 

Cavates were brought to the attention of 
the profession in other contexts in the early 
twentieth century. J. W. Fewkes made several 
references to Pajarito cavates in his presidential 
address to the Anthropological Society of 
Washington and included some photographs of 
cavates (Fewkes 1910). W. B. Douglass (1917) 
presented several sketch plans and sections of 
cavates in a paper for the nineteenth Congress of 
Americanists. His profiles show several cavate 
features, some quite common and some, such as 
benches and altars, rare. 

Hendron (1940, 1943) did some of the 
most carefully recorded work in cavates. He 
excavated five masonry rooms and the four 
cavates associated with them in the center of 
Group M, above the NPS Residence Area. His 
intention was to thoroughly examine some 
structures in order to be better able to stabilize 
them. The rooms contained roofing material, 
cow bone, and Tewa blackware, indicating 
historic (ca. 1680 Pueblo Revolt) use of these 
particular structures. The uppermost walls were 
underlain by other wall alignments, some of 
which may have been used as foundations for 
the later walls. Hendron was very impressed by 
the fragility of the natural canyon walls, which 
led him to emphasize the danger of living in 



cavate rooms and to suggest that the walls 
probably receded very rapidly. He believed that 
the rooms visible at that time may have been 
preceded by earlier rooms later obliterated by 
erosion. Although the Bandelier Tuff is clearly 
a very soft material, it seems probable that 
Hendron overestimated the rate of cutting (see 
Carlson and Kohler 1989:59). Forty-five years 
after he worked there, the structures showed 
relatively little change. Hendron estimated the 
period of use of the sites as 400-500 years. He 
described firepits, plaster "dados," smoke vents, 
a basalt threshold, and depressions in the floor, 
which he said suggest sleeping spaces. Given 
the smoke blackening and the method of 
ventilation, Hendron was inclined to doubt that 
the cave rooms were used for habitation. He 
made some extremely nice drawings and sections 
(now in the park archives), but said little about 
recovery of cultural material. The rooms he 
stabilized are clearly visible today. Maxon 
(1969) reported a single tree-ring date of a.d. 
1493 from Group M. 

In his master's thesis, J. F. Turney 
(1948) wrote up the artifacts from Hendron's 
excavation as well as some further material from 
the excavation of a drainage trench in Group M. 
He noted that "it has been necessary to remove 
this material from its unfinished status and bring 
it to a conclusion as an aid to further research" 
(1948:i). The majority of the thesis is devoted 
to a discussion of pottery classification and 
description of types; the types identified by 
Turney are listed in Table 1.1. In connection 
with the present study, surface sherd counts 
from the same area were conducted and are 
presented in Table 2.3. 

The fauna! material includes deer, bison, 
turkey, and bear bone and a few worked pieces. 
Manos, metates, and axes were recovered, as 
well as bifaces and at least 10 projectile points. 
Obsidian is the most abundant chipped stone 
material (though Turney stated that "obsidian is 
brittle and not too well-suited for chipping" 
[1948:64]). Though the perishable materials 



10 CAVATE STRUCTURES 



Table 1.1. Group M Ceramics Reported by Turney (1948) with Counts from Cavate Surface Material 
Analysis (See Table 2.3). 



Ceramic Type 


Turney 


Group M Sample 


Frijoles Sample 


Santa Fe Black-on-white 


57 


5 


16 


Wiyo Black-on-white 


24 


1 


3 


Abiquiu Black-on-gray 


30 


- 


- 


Bandelier Black-on-gray 


a 


26 


62 b 


Sankawi Black-on-cream 


56 








Tewa Polychrome 


61 


- 


- 


Agua Fria Glaze-on-red 


2 


4 


9 


Cieneguilla Glaze-on-yellow 


5 


1 


1 


Glaze B yellow 


1 


1 


2 


Glaze C 


5 


1 


5 


Glaze D 


11 


7 


34 


Glaze E 


15 


3 


7 


Glaze F 


5 


- 


- 


Zia Glaze 


4 


- 


- 


Early glaze 


32 


- 


- 


Middle glaze 


203 


- 


- 


Late glaze 


55 


- 


- 


Zia Polychrome 


13 


- 


- 


Kapo polished blackware 


73 


1 


1 


Culinary 


c 


469 


1144 


Total 


652 


519 


1284 



•Not given; presumably abundant: shown as "major occupation." 

b Biscuit B. 

c "By far the largest type"; "culinary ware is of little value" (1948:47). 



INTRODUCTION 



11 



seem to have come from the masonry rooms 
rather than the cavates, they are abundant and 
remarkably well preserved. The inventory 
includes a digging stick, two weaving tools, two 
bow fragments, several arrow shafts, "a meager 
three" fire hearths, some carved sticks, a cradle 
board, and other wood attributed to craft 
wastage. Basketry, cordage, yucca fiber, 
feathers, a feather blanket, and a piece of 
woolen textile were also present. Turney listed 
24 pieces of leather, including moccasin 
fragments, sewn buckskin, and a thong. A 
small bowl of tobacco was found, analyzed, and 
partially smoked(!) by Hendron (Hendron 1946). 
Also recovered were corn plant parts (bundles of 
leaves, more than 600 cobs) and cucurbit 
peduncles and rinds. Following Mera (1932, 
1934) and presumably Hendron, Turney 
concludes that a reoccupation, possibly during 
the Pueblo Revolt, was probable. The types of 
perishable materials and the degree of 
preservation indicate that they may relate to the 
later occupation suggested by some of the 
pottery types. 

In 1939 and 1940, R. H. Lister 
conducted extensive stabilization work in 
Bandelier. Much of his time was spent at Long 
House (Lister 1939), but he also worked at 
Otowi and Tsankawi, west of Camp Hamilton, 
in Pueblo Canyon, and along the Rito de los 
Frijoles. In the area outside Frijoles he located 
567 caves and worked on more than half of 
them. Most of his work consisted of building 
dams at the entrances of the cavates or 
diversions above them to prevent water from 
running through them. In Frijoles Canyon he 
did similar preventive maintenance and 
rechinked some masonry walls. Lister's dams 
are difficult to see today, but they are probably 
doing their job since water damage seems to be 
less a problem than human impact. The 
stabilization records for Long House deal only 
with masonry walls. The records for the work 
in "caves" (Lister 1940a,b) contain many before- 
and-after photographs of the cavates in which 
Lister worked (see Chapter 2), as well as brief 



descriptions of the tasks performed. The work 
done outside Frijoles (Lister 1940b) was 
concerned almost entirely with landform and 
drainage, though Lister did note and photograph 
some rock art and disperse some rock corrals in 
front of cavates. 

The site group west of Camp Hamilton 
has an especially high density of cavates; 
Lister's map shows 161 in less than half a 
kilometer. Lister assigned numbers to cavates in 
each of the areas where he worked. In each 
area the numbers start with one, and C is used 
as a prefix in all areas (e.g., C35). He prepared 
a map for each area showing the configuration 
of the cliff. Lister's (1940a) descriptions of 
cavates in Frijoles Canyon are somewhat more 
detailed than those for the other areas, and the 
work he performed included some masonry 
pointing. The most common modification he 
made involved rearranging fill in front of 
cavates, but Lister makes no mention of artifacts 
in or outside of them. He summarized his 
cavate work in several notes in the Southwestern 
National Monuments newsletter. 

In 1960 C. Johnson, a graduate student 
at the University of New Mexico, removed a 
secondary burial of a child from "Cave Room 
C54 Tsankawi Ruin." This room is located in 
the upper, gray tuff cliff, just below the main 
pueblo, and is part of the group our crew 
recorded at Tsankawi (LA 50976). Matting, 
cordage, a few sherds and lithics, and a 
Bandelier Black-on-gray bowl were present in 
the back corner of the room, under about 8 cm 
of fill. Since small bones were missing and the 
long bones were stacked below the cranium, this 
was quite clearly a reinterment. A floor pit, 
apparently unrelated to the burial, was beneath 
it (Johnson 1960). 

In 1962 James Maxon, the park 
archaeologist at Bandelier, cleared the floor of a 
cavate in Mortandad Canyon, then in the Otowi 
Section of the monument (Maxon 1962). This is 
a fully enclosed, 3 x 3.6 m cavate with two 



12 CAVATE STRUCTURES 



doors, one blocked up; a firepit near the blocked 
door; four large niches; and abundant rock art. 
Because this room is the largest in its group, 
Maxon called it a kiva. Some corn remains and 
a few sherds were recovered from the 
excavation; the decorated sherds from the room 
and the area outside are Santa Fe Black-on-white 
and Wiyo Black-on-white. The petroglyphs in 
this cavate largely defined Steen's Mortandad 
style of rock art (Steen 1979). 

Maxon analyzed material from two Los 
Alamos Archaeological Society excavations for 
his master's thesis (1969). One of the sites was 
the Tshirege Cave Site, consisting of 13 
masonry rooms, 6 cave rooms, and a court area. 
Tshirege is located near the modern community 
of White Rock and is thus not far from 
Tsankawi. Like Tsankawi, Tshirege is a large 
masonry pueblo with cavates below it, including 
the "Tschirege Cave Site" (Maxon's spelling). 
Maxon stated that: 

available photographs indicate that these 
cave rooms were also typical of other 
cave rooms in the area. The rooms 
rarely exceed 10 feet square in size and 
ceilings are rarely more than 6 feet in 
height. Occasionally cave chambers 
which were used as ceremonial rooms or 
kivas were somewhat larger. . . Usually 
[they are] behind surface dwellings . . . 
rarely used alone. 

Cave rooms typically had adobe 
plastered floors, often showing several 
layers of plaster. The walls were also 
plastered to a height of about 30 inches. 
The plaster was often colored red or 
white. Above the plaster the walls are 
heavily soot coated from fires, either in 
the caves themselves or from adjoining 
rooms to the outside. In addition to the 
doorways, sometimes a small ventilator 
hole was located near the ceiling of the 
cave. The cave rooms were often 
interconnected... Apparently cave rooms 



were not excavated more than one row 
deep into the cliffs. 

Aside from their good insulation 
from both heat and cold, the caves had 
little to offer for day to day living. The 
lack of light, ventilation, and their 
generally small size made them less 
desirable than the rooms built in front. 
Nevertheless, evidence of use of fire, 
storage niches, and repeated refurbishing 
of the floors suggest that the caves were 
utilized as much as the outside rooms 
(Maxon 1969:49-50). 

The ceramics suggest that this part of 
Tshirege was used from the late fourteenth 
through the early sixteenth century. Material 
consisted of fairly abundant pottery, some 
milling stones, mauls, and chipped stone 
dominated by obsidian. The only perishable 
mentioned from the collections is some leather. 

Charlie Steen (1977, 1982) has 
assembled information on work done over many 
years on the lands of Los Alamos National 
Laboratory. Although he made "no particular 
effort ... to locate groups of cavate rooms" 
(1977:3), he made several observations 
concerning cavates. He believes they relate to 
the period on the Pajarito Plateau when the 
larger sites were being occupied. He argues that 
cavates served primarily for storage and for 
ceremony. The argument for ceremony rather 
than habitation is based on several contentions: 
that cavates are intentionally smoke blackened 
rather than blackened by heating fires 
(replastered walls are sooty but not black), that 
hearths near doors are kiva features, and that 
cavates often have artwork in them (1977:15- 
17). He believes the very small, blackened 
examples were used for individual meditation. 
The larger blackened and plastered cavate rooms 
with rock art "served as religious or ceremonial 
rooms for basic families, and each was similar 
to a family chapel" (Steen 1979:42). Steen also 
defined a rock art style, the Mortandad style 



INTRODUCTION 13 



mentioned above, which he says is limited to a 
small subarea of the Pajarito Plateau and found 
strictly in cavates. He attributes it to the late 
fourteenth century. The style is characterized by 
incisions into blackened tuff and by the presence 
of Awanyus, birds, Kokopelli, and an 
anthropomorphic figure that Steen likens to the 
Toltec sun god. Steen's viewpoint concerning 
the Pajarito Plateau cavates found a wide 
audience in a 1982 issue of National 
Geographic, which has long been an outlet for 
discussions of cavates (Beam 1906; Canby 
1982:578-579, 592). 

The Pajarito Field House Project~a part 
of the Pajarito Archaeological Research Project 
(PARP) of the University of California, Los 
Angeles-excavated the shallow deposits in an 
isolated cavate (LA 52333) in the area south of 
Puye (Preucel 1985, 1986a; Hill and Trierweiler 
1986). This single room contained three 
hearths. Both squash and corn remains were 
recovered. The few sherds associated with it 
and those on the talus in front were Santa Fe 
Black-on-white and Tesuque corrugated. Robert 
Preucel suggested that because of its isolation 
this cavate could have functioned as a field 
house. As part of the PARP, Justin Hyland 
collected data on features and room dimensions 
for a total of 44 cavates in 2 adjacent groups in 
Garcia Canyon, between Tsankawi and Puye. 
He presented the results of this recording and its 
analysis in his honors thesis (Hyland 1986). 
Although some of the variables measured were 
different, and the means of measurement also 
differed, Hyland's study is by far the most 
nearly comparable to the present one, and it 
monitored many of the same attributes. In 
addition to the cavates recorded by Hyland, 
PARP recorded the locations and some 
measurements for 431 other cavates in a total of 
35 locations (R. Preucel, personal 
communication, 1988). The PARP survey also 
documented several large, Late Coalition 
pueblos with extensive associated "cavate 



villages" in the area south of Puye (Preucel 
1986b: 8; 1987). 

As part of an excavation program related 
to the Bandelier survey, a Washington State 
University field school under the direction of 
Timothy Kohler excavated a single cavate 
chamber in Frijoles Group M near the rooms 
excavated by Hendron and outside the area 
recorded by our crew. The fill of this room was 
primarily disintegrated tuff, with sloughed wall 
plaster near the floor. Materials recovered were 
similar to those we observed during recording: 
a few cultivar remains and very sparse artifacts. 
Seven features were recorded, of which the three 
floor features-two bins and a cist-were rare or 
absent in our sample (in which there are 
relatively few floor features). A deep, heavily 
modified niche is unlike any recorded in this 
study (Carlson and Kohler 1989). 

Cavates have long attracted attention, 
and it seems that in the past hundred years they 
have been found to contain little material. That 
situation, however, may well relate as much to 
their visibility as to the condition in which they 
were left (in some senses they have never been 
completely "abandoned"). Although most of the 
people who have studied cavates have noted 
replasterings, smoke blackening, and domestic 
features, they generally agree that for reasons of 
space and ventilation, cavates would be an 
undesirable place to live. Three functional 
categories have been defined: small, unsmoked 
rooms used for storage; rooms with smoke 
blackening and other features that may have 
been habitations; and the largest cavates, smoke- 
blackened and containing rock art and sometimes 
loom anchors, which have been called "kivas." 
Although the last category in particular is loaded 
with assumptions and inferential leaps, these 
categories may at least be tested with detailed 
data from cavates. The PARP study and the 
sample reported here are a beginning, but 
expanded research is needed if we are to make 



14 CAVATE STRUCTURES 



meaningful regional statements about function 
and variability. 

The Present Project 

In the summer of 1986 a crew of three 
National Park Service (NPS) archaeologists, 
assisted by four volunteers on varying schedules, 
spent six weeks recording measurements and 
features of cavate rooms in Frijoles Canyon and 
in the Tsankawi section of Bandelier National 
Monument. The locations of cavate features in 
Frijoles Canyon and at Tsankawi have been 
known for a long time (e.g., Hewett 1909a; 
Lister 1940b). This project recorded data on 
condition, measurements, associations, and 
visible features for a sizable number of cavate 
rooms. These data were collected as 
consistently as possible, recorded in a coded 
format, and entered into a computer data base so 
that they could be readily manipulated and easily 
recalled. The project produced location sketches 
and maps, along with photographs of feature 
openings and selected features, and videotapes of 
all rooms recorded. Immediately following the 
fieldwork, we entered the data into the computer 
and prepared a preliminary report, but we lacked 
funding and personnel for either checking or 
analyzing the data. A later contract permitted us 
to check for errors in coding and data entry and 
to prepare baseline data on cavate feature 
occurrence and measurement. This study is the 
result of the latter analysis combined with the 
preliminary report. In addition to the foregoing 



oudine of previous archaeological work done in 
cavates, it describes the sites in which recording 
was done, discusses the recording procedures 
followed, and presents the results of descriptive 
and interpretive analyses using the data gathered. 
Beyond the goals of better cavate management 
and description, several higher level questions 
motivated this study. 

Several interesting questions arise from 
the project. Can we group cavates based on size 
and features present, and can we interpret the 
possible functions of these groups? Do certain 
types of cavate tend to occur in certain 
locations? How do Tsankawi cavates differ 
from those in Frijoles, and can we attribute 
those differences to the cultural boundary 
between the Tewa and the Keres, which those 
groups traditionally consider to have existed 
between those areas (e.g., Hewett 1938; Steen 
1977)? The data assembled here allow us to 
begin discussing these questions; answering them 
will require following them much further. 

In capsule, then, this project was 
designed to establish: (1) an estimate of the 
ranges of variability and covariation of attributes 
in cavates; (2) a procedure for recording these 
features; (3) the current condition of a sample of 
structures as a baseline for future maintenance 
and for the types of disintegration that occur in 
cavates; and (4) the coherence and logic of 
various interpretations of cavates. 



Context, Descriptions, and Chronology 



Edgar Hewett (1938:27, 34) takes credit 
for naming the east flank of the Jemez 
Mountains the Pajarito ("little bird") Plateau, in 
spite of his peculiar statement that "the country 
is almost devoid of birds" (Hewett 1938:30-31). 
As he defined it, the plateau extends from the 
Chama River on the north to Canada de Cochiti 
on the south and is bounded by the Rio Grande 
on the east. The name comes from Pajarito 
Canyon, which in turn takes its name from the 
translation of the Tewa word Tshirege 
(Harrington 1916:282-283; Lange et al. 
1975:58, 77; Hendron 1946:89; also spelled 
"Tsirege," "Tschirege," and "Tzirege"), a large, 
Classic period pueblo ruin on the outskirts of 
modern White Rock. 

The Pajarito Plateau is a place of great 
drama and beauty. Its geological history 
culminates in a huge explosion; its great 
elevational range and resulting moisture give rise 
to biological diversity and splendid panoramas; 
its human history is long and highly varied. 
This study examines a small spatial and temporal 
portion of this broader context. In addition to 
summarizing the setting, this chapter describes 
the areas in which we worked and how they 
were selected. It concludes with a discussion of 
the means and problems of the chronological 
placement of cavates. 

The Setting 

Geology and Environment 

Even more than in most prehistoric 



settings, the geology of the Pajarito Plateau was 
of critical importance to how its inhabitants 
adapted to life there. The massive, relatively 
easily excavated tuff deposits exposed by the 
canyons of the plateau made possible the 
construction of cavate dwellings. The source of 
this tuff was a major geological event during the 
Pleistocene: the explosive eruption of the Valles 
Caldera followed by an ash flow that spewed 
forth a couple of hundred cubic kilometers of 
volcanic ash (Ross et al. 1961:141; Heiken 
1979; Mathien et al. 1993). The singularity of 
the event accounts in large part for the limited 
occurrence of cavate dwellings. The Jemez 
Mountains were formed by millions of years of 
volcanism, but it was the "climactic and terminal 
stage" that formed the Tewa Group of tuffs and 
lavas in which the cavates were constructed 
(Figure 1.1; Bailey et al. 1969:12-15). The 
Tewa Group contains both the Valles Rhyolite 
and the Bandelier Tuff; the Bandelier Tuff is the 
formation into which the cavates were carved. 
The caldera left by the eruption, the ash flow 
blanket and its later dissection, are the main 
features of the landscape occupied by the Pueblo 
peoples in the twelfth through sixteenth 
centuries. 

The Jemez Mountains rise to 3500 m 
(11,500 ft.). They form an effective moisture 
trap, and several streams flow out of them, 
including the Rito de los Frijoles. The presence 
of water, the sharp elevational differences, and 
the softness of the tuff have predictably led to 
dissection of the Pajarito Plateau by numerous 
deep, often sheer-walled canyons radiating from 



15 



16 CAVATE STRUCTURES 



the caldera. The Bandelier Tuff has been 
divided into two members-the Otowi being 
overlain by the Tshirege--each of which consists 
of ash flow units resting on thinner pumice beds 
(Bailey et al. 1969:12-14). Cavates occur 
primarily in the Tshirege (upper) member, 
which consists of "a series of cliff-forming 
welded ash flows" (Bailey et al. 1969:13; Kelley 
et al. 1961:56-59). The lowest cavates recorded 
at Tsankawi are cut into a distinctive reddish tuff 
that is much softer than the overlying gray tuff 
and contains larger chunks of pumice. In this 
location Bailey and others (1969) describe the 
Tsankawi Pumice Bed at the base of the upper 
member of the Bandelier Tuff. The layer into 
which the cavates are cut, however, is thicker 
than their description of the pumice bed (they 
measure it at around 1 m). In any case, these 
cavates are also in the Tshirege member of the 
Bandelier Tuff, very near its base. 

Grant Heiken (1979:569) notes that 
deposits at the base of pyroclastic flows are 
easier to excavate. In response to an inquiry 
about patina formation and, more specifically, 
the variability in the tuff at Tsankawi, he wrote 
the following: 

Most of the variations in the tuff at 
Tsankawi ... are related to 
postdepositional processes. Obvious 
facies within the tuff, which are soft and 
easy to excavate, include nonwelded 
bases and pumice falls. Harder tuffs 
include those cemented by secondary 
minerals in the vapor phase zone (that 
zone near the top of the composite 
section of pyroclastic flows affected by 
hot gases rising through the cooling tuff 
deposit) and the thin, resistant layers 
cemented by zeolite cement. The latter 
were located at the top of the ground 
water table that was present before the 
canyons were excavated by erosion. 
Multiple resistant beds can represent a 
record of declining ground water as the 
canyons were growing deeper and wider 
with time. The flat benches at Tsankawi 



are tops of the more resistant zeolite- 
cemented tuff. 

Nearly all cavates are located within 
distal regions of the pyroclastic flows 
where the tuff is nonwelded. Closer to 
the source, for example above 
Ponderosa Campground, the tuffs are 
welded; these would have been 
impossible to excavate, being hard and 
dense. (G. Heiken, personal communi- 
cation, 1986) 

The well-watered uplands contain lush 
vegetation and associated montane fauna 
(Mathien et al. 1993:6-8). Because the elevation 
drops rapidly to the Rio Grande at around 1600 
m, there is considerable biotic diversity in a 
fairly small area. Alpine tundra, spruce-fir- 
aspen, ponderosa-oak, pinon-juniper, and 
riverine plant associations are all found within 
25 km of the study areas, depending on 
exposure and elevation (see Powers 1988 for 
more detailed discussion of environmental 
zones). Even in the lower elevations of the 
plateau, the growing season varies considerably, 
from 120 to 180 days (Hubbell and Traylor 
1982:29). 

Cultural Context 

For an area seemingly well suited to 
supporting populations subsisting by hunting and 
gathering, remarkably few sites of the Archaic 
and early parts of the Pueblo eras are known on 
the Pajarito Plateau. David Stuart and Rory 
Gauthier (1981:48-49) found that fewer than 12 
percent of all components in the state survey 
files date to before a.d. 1 175. More sites from 
the earlier period have been and will be found as 
more comprehensive archaeological work takes 
place, such as the preceramic pithouse near 
Otowi (Lent 1988). Present samples, however, 
indicate relatively low human population and use 
in the area in early prehistory. This situation 
changed dramatically, however, at around 1175. 
At that time there was a sudden profusion of 
small pueblos characterized by rectilinear room 
blocks, pit structures, and ceramics dominated 



CONTEXT 17 



by Santa Fe Black-on-white: the Coalition Period 
(Stuart and Gauthier 1981:45-51; Cordell 
1979:53-64; Preucel 1987; Powers 1988; 
Mathien et al. 1993:9-34). Judging from 
ceramics, it is probable that the first cavate 
structures were dug sometime during the 
Coalition Period (see below). 

Following the Coalition Period, between 
about 1325 and 1540 the population of the area 
became increasingly aggregated, as seen in the 
archaeological record for much of the northern 
Rio Grande, including the Pajarito Plateau 
(Kohler and Linse 1993:3-5). At this time~the 
Rio Grande Classic-there were fewer but much 
larger settlements. Large, free-standing pueblos 
were built, such as Tyuonyi and Tsankawi, and 
on the Pajarito Plateau the large pueblos have 
large groups of cavates nearby. Other, well 
known sites having both cavate clusters and 
large, free-standing structures include Otowi, 
Navawi, Tshirege, and Puye. These associations 
and the predominance of ceramics, such as 
Bandelier Black-on-gray and Tsankawi Black-on- 
cream, on cavate sites suggest that at least 
clustered cavates are a part of the Classic Period 
aggregation. In the historic era, following the 
Classic Period, permanent habitation on the 
Pajarito Plateau shrank dramatically, and almost 
all Pueblo settlements were located in the major 
river valleys. The plateau remained an 
important subsistence and sacred resource area 
for the Pueblo peoples, as well as a refuge, but 
its time as a location for large human 
populations was over until the coming of the 
nuclear age to Los Alamos. 

The Sites 

Cavate Groups 

Both Frijoles Canyon and the area 
around the main ruin at Tsankawi were very 
densely settled, so that the concept of discrete 
sites in these areas is somewhat suspect. Hewett 
and Chapman divided the cavates in Frijoles into 
Groups A through M based on breaks between 
clusters, which are often caused by drainages or 



stretches of cliff unsuitable for cavate 
construction. The separation between the groups 
they defined is only a few meters in several 
cases, and the groups vary considerably in size. 
Whether or not this long string of cavates was 
14 or more settlements, as implied by this 
topographic grouping, can only be inferred by 
careful study. Our recording is a early step in 
making this inference. 

In 1986 we spent time recording in five 
groups of cavates (Figures 2.1, 2.2). The 
Hewett-Chapman groups remain useful to the 
NPS as a framework for management, and we 
used them as the first stratum for our recording 
sample. Because of short time and small crew 
size, we recorded all of one of the Hewett 
groups and parts of three others. In addition, 
we worked at a fifth group in the Tsankawi 
portion of Bandelier National Monument, 11 
straight-line km northeast of the cavates in 
Frijoles Canyon (Figure 1.1). 

The sample of cavates selected was 
designed to assess several dimensions of 
variability. Within Frijoles Canyon we were 
interested in whether a number of locational 
variables influenced cavate morphology: 
upstream or downstream location within the 
canyon; vertical and horizontal proximity to the 
Rito; size of cavate group; location within a 
group. At the next level, we were interested in 
differences and similarities between the Frijoles 
cavates and a group of cavates outside Frijoles 
Canyon (the Tsankawi sample). In Frijoles, 
dimensions of locational variability are to some 
degree correlated; that is, cavates in the upper 
end of the canyon are closer to the Rito because 
of the canyon topography and the intersection of 
the stream with the tuff strata. Tsankawi Mesa 
has at least three major tuff types, and we 
studied some rooms from each. 

The rationale for recording a stratified 
sample of groups rather than a random sample is 
grounded on both practicality and information 
yield. First, to accurately sample the whole 
Frijoles or Tsankawi population would require 



18 CAVATE STRUCTURES 




Cliffs 



Studied Cavates 



GROUP M 



500 meters 



RAINBOW HOUSE 



2000 feet 

I 



Figure 2.1. Map ofFrijoles Canyon showing the location of the cavate groups discussed in this study 
and other major sites. Only selected contours from the USGS Frijoles Quadrangle are 
shown, with the cliffs of the north side and the cliffs and steep slopes of the south side 
shown by shading. 



CONTEXT 19 




Figure 2.2 Map of the Tsankawi section ofBandelier National Monument, showing the main pueblo 
of Tsankawi. LA 50976 and three groups ofcavates mapped by Lister are scattered along 
the south edge of the mesa. By request ofBandelier National Monument, the locations 
of these fragile cavate groups are not shown. Selected contours and features are taken 
from the USGS White Rock Quadrangle. 



20 CAVATE STRUCTURES 



complete inventories of each, and those 
inventories did not exist in 1986. Second, 
locating randomly selected features would be 
time-consuming and it would be difficult to 
determine which other features should be 
included. Third and more important, there are 
good archaeological reasons for recording 
groups. It is extremely unlikely that any single 
cavate in either of these locations was a site unto 
itself in terms of prehistoric use. By recording 
groups, it is possible to gain some idea of 
whether size or functional groups are associated 
in a regular way. It might be possible to infer, 
for example, whether or not smaller groups of 
cavates can be considered individual-use units 
(see Figures 2.3, 2.4.). 

Through an unfortunate oversight, the 
field crew was given an incorrect set of 
Laboratory of Anthropology (LA) site numbers 
for use in the field. These numbers were used 
on all forms, photo records, and notes, and on 
photo boards. Although the correct, official 
numbers are used throughout this text, the 
presence of the field numbers in so many places 
requires a concordance between official numbers 
and field numbers: 



Group 


Official No. 


Field No. 


M 


LA 50972 


LA 50020 


A 


LA 50973 


LA 50021 


I 


LA 50974 


LA 50022 


F 


LA 50975 


LA 50023 


Tsankawi 


LA 50976 


LA 50024 



The field numbers have been retained in the 
computer data base for purposes of matching 
field records and photographs with final records. 
A plan map and elevation or profile sketch is 
included for each study group in the following 
descriptions. The variability in these figures is 
the result of three different recording techniques. 
During the 1986 fieldwork each Frijoles group 
was mapped using tapes and compass. Cavates 
in the Tsankawi group were placed on Lister's 
map. Rough field elevations were drawn 
showing locations, and then drawings were made 
on Mylar overlays on photographs. Groups F, 



I, and M in Frijoles were included in the later 
sample survey of the monument. During 
recording for the survey, plane table maps and 
careful elevations were made. When available, 
the more detailed maps and drawings from the 
park survey have been included here. When 
these were unavailable, drafted versions of the 
1986 recording have been used. 

Frijoles Group A, LA 50973 
(Held no. LA 50021) 

Group A is located in a lovely, park-like 
wide spot in the canyon. It is a large group, 
running a couple of hundred meters along the 
base of the cliff and consisting of around 130 
cliff-associated rooms and several substantial 
rubble mounds representing masonry room 
blocks (Figures 2.5, 2.6). Group A is separated 
from Group B by a projection in the cliff, but 
the two are quite close together. In spite of its 
designation, which would seem to indicate that 
it is the first group of cavates in the canyon, 
there is a substantial cluster of cavates 1.1 km 
upstream from Group A (see the description of 
Cuevitas Arribas, below). Group A is located at 
the upper end of the fairly continuous string of 
locations with cavates that stretches for about 2 
km through the central occupation area of 
Frijoles Canyon. Much of this distribution may 
be explained by canyon width: Group A is 
located in the last wide spot in the canyon as one 
proceeds upstream, the upper end of the wide 
part of the canyon where the Rito is not deeply 
entrenched. Like Long House (Group D), 
Group A is located very near the canyon floor 
rather than at the top of a talus slope, the 
location of most of the other Frijoles cavates. 

The U.S. Geological Survey (USGS) 
quadrangle indicates that the creek is at an 
elevation of 6150 ft (1875 m) where it passes 
Group A. Altimeter readings at the cavates 
indicate a difference of 150 ft (46 m) from the 
Visitor Center benchmark, or about 6210-6220 
ft (1893-1896 m). The cavates of Group A are 
all at about the same elevation because the base 
of the cliff is at the same level for most of the 



CONTEXT 21 




Figure 2.3. View of Group I from across the canyon. Masonry rubble is 
visible in front of the concentration of cavate rooms. The 
relatively recent rockfall on the righthand side of the group 
partially covers rooms 30-34. The group extends from the 
lower chamber on the left and there are two rooms out of the 
frame to the right. 




-* 1 



•^fe^ 








Figure 2.4. Upper Group M viewed from across Frijoles Canyon. The 
area recorded extends from out of the picture on the left to 
the small tent rocks at the far right. Note both the extensive 
rubble and the heavy use of cavate chambers that had 
masonry closings. 



(63)65(66) 



^Ue, -' 4 167^8/J^ 



CONTEXT 23 



73 ^ 





Ceramic Sample Area 
Extends 25 Meters — 



LA 50973 

GROUP A PLAN VIEW 



I 44 J 

i i Upper Story Cavate 
tefe;zP~ 1st Story Cavate 



fl(n ^ Extent of Rubble 
>$£&, Bed Rock 
^ Boulder 



10 



i i i i i i_ 



15 

_1_ 



20 



10 

_1_ 



m«t»t» 
20 30 



40 



50 



ip A sample. Room numbers between dotted lines are 
I at the bottom of the stacked room numbers. Back rooms 
Ing a compass and tapes in 1986. 



CONTEXT 23 



(63)65(66) ^ 73 » 

\ / 62 / I | ' ,'70 '71 \ 7 ?X 

\ 59,60, \ / .1 -' ! 67 ' 




Ceramic Sample Area 
Extends 25 Meters — i 



LIMIT OF CERAMIC SAMPLE AREA 



LA 50973 

GROUP A PLAN VIEW 



43. 



Upper Story Cavate 



l?feiT~ 1st Story Cavate 
"rmir^ £"'•"• of Rubble 
■'?t>.: Bed Rock 
(7/X Boulder 



10 



i — i — i — i i i_ 



10 20 30 40 



Figure 2.5. Room plan view for the Group A sample. Room numbers between dotted lines are 
arranged with lowest story room at the bottom of the stacked room numbers. Back rooms 
are in parentheses. Drawn using a compass and tapes in 1986. 



Group at Back of 



CONTEXT 25 




61 63 | 66 
9 64 



sample. Drawn from field sketches and photos in 1986: 
distance from Cavate 1 to Cavate 73 is approximately 



CONTEXT 25 



Group at Back of Cleft (same level as 1-5) 



LA 50973 

GROUP A PROFILE 

sags Wall Alignment 
••• Viga Holes 



10 

i 



meters 
20 30 

_l L_ 



40 

I 




20 

_l 



51 54 ), 56 




£H 



Figure 2. 6. Elevation sketch for the Group A sample. Drawn from field sketches and photos in 1986: 
scale approximate (straight-line distance from Cavate 1 to Cavate 73 is approximately 
93 m). 



CONTEXT 27 



group. The USGS contours are quite confusing 
here. While converging contours are to be 
expected given the size of the cliffs, the fashion 
in which they converge here is enigmatic. 
Group A, as noted, is located in a relatively 
wide spot in the canyon. This fact is suggested 
by the contours, but the 6200 and 6300 foot 
lines converge just below A and show a much 
broader bench than appears to exist 200 ft (60 
m) above the Rito. While the cavate symbols 
seem to be properly placed horizontally (there 
are three for all of Group A), the contours seem 
wrong, and the altimeter supports our 
disagreement with the contours. Therefore, we 
chose 6210 ft (1893 m) as an elevation for 
Group A and estimated the variations from that 
baseline. 

The tuff at the base of the cliff is more 
variable in a shorter vertical section in the area 
of Group A than in any other area in which we 
worked. There are several layers that differ in 
color, texture, and durability, including a 
crumbly gray unit containing quantities of black 
rock, a coarse-grained but fairly compact white 
layer, a brownish layer, and the gray-white finer 
texture seen commonly elsewhere in Frijoles 
cavates. All of these units intersect the 
structural parts of Group A. 

The portion of Group A that we 
recorded extends from the uppermost rooms 
about halfway through the group. The upper 
end of the group is quite well defined. There is 
a steep, deep cut in the canyon wall at the upper 
end of the wide area in the canyon, which 
contains several pockets well up the cliff. These 
could have contained rooms but do not appear to 
have had any, and no further rooms seem to 
exist between upper Group A and Ceremonial 
Cave (however, a thorough examination has not 
been made). The lower end of our study in 
Group A was defined more by the exigency of 
field time than by natural breaks within the 
group. Since the sample of Group A was small 
relative to the group and we had an extra day 
and a half at the end of the field season, the 



naturally defined subgroup extending to A-36 
was expanded to A-73. 

Our count of 130 rooms showing in 
some way on the cliff includes 40 rooms 
recorded in detail in July 1986 and 90 rooms 
below that stopping point. (The latter figure is 
based on three counts giving 90, 92, and 81; the 
count of 81 was made from a point well away 
from the cliff, from which vantage point many 
low rooms are not easily visible.) Usually the 
process of detailed recording revealed more 
rooms than were recorded with careful but 
necessarily more superficial counts, and this has 
a minor effect on the relative positions recorded 
for some cavates. The overall area of the rooms 
recorded is about 95 m along the cliff by 20 m, 
allowing for recesses in the cliff and rubble 
areas. It is not entirely clear whether the base 
of the cliff in Group A had a continuous 
structure along it or whether there were 
separations between structures. Rows of viga 
holes directly above one another show that major 
remodeling did take place in the masonry rooms 
of Group A, which makes determination of the 
presence of continuous built structure even more 
difficult. Whether or not the building was at 
some time continuous, most of the cliff did have 
some structure placed against it. The amount of 
building rubble now visible varies considerably 
along the base of the cliff, with greater 
quantities present in recessed areas in the cliff. 
The total number of rooms is probably more 
than twice the number visible on the cliff. 
Considered as a site, then, Group A was 
probably at least as big~and as long~as Long 
House and was probably comparable to Group 
M. Particularly at the lower end, but elsewhere 
as well, much of the site was multiple stories; in 
some places it was at least two, and quite often 
four. Its location near the canyon floor probably 
made it easier to erect taller structures here than 
on higher, steeper talus slopes at other cavate 
groups. 

Several other archaeological features are 
present in the vicinity of Group A. The most 
notable is a masonry-lined kiva below the 



28 CAVATE STRUCTURES 



approximate center of the group. This structure 
is shown on the Hewett-Chapman map and 
remains distinctly visible, which indicates that it 
has probably been at least partially excavated, 
though Park Archaeologist Bill Sweetland says 
there is no record of such an excavation 
(personal communication, 1986). Visible 
structural remains are otherwise scarce on the 
relatively flat canyon bottom below the group. 
This absence of structures is probably related to 
the conservation of watered bottomland for 
crops, though the presence of a low terrace 
would have required a substantial ditch to 
irrigate much of this part of the canyon; pot 
irrigation would certainly be easier here than in 
most of the Southwest. On the first ledge above 
the base of the cliff are several petroglyphs. We 
did not venture up, but the ledge looks large 
enough to have had some structures on it. Two 
constructed routes seem to extend at least as far 
up as the ledge at the upper end of the room 
group (vicinity of rooms A-9 and A-10). Routes 
to the canyon rim above this ledge as well as 
slightly up-canyon are suspected, but none of 
these has been actually climbed. The only other 
known recorded archaeological work at Group A 
is the repairs made by Lister in 1939 (Lister 
1940a). 

Considering the size of Group A, there 
seems to be remarkably little trash present on 
the surface. Some of the other groups have 
steep slopes below them, which may accelerate 
removal of trash by washing, but this is less a 
problem at Group A. Based on casual 
observation and McKenna's more systematic 
study, Group M has more trash on the slope 
below it than does Group A. Lower visibility 
due to the denser vegetation of Group A is 
probably one reason for this, but a more 
important one is its heavier tourist traffic. 
Excavated tests of trash quantities (and qualities) 
at any of the cavate groups would be of 
considerable interest. 

This area has been closed to visitation 
for some time (Sweetland, personal 
communication, 1986). Nonetheless, Group A 



seems to have more graffiti than any of the other 
groups in which we worked, including 
Tsankawi, which receives heavier, unsupervised 
visitation. Group A clearly has visitors and has 
had them for a long time. Our presence there 
may have reduced illicit visits, but in the five or 
so days we spent recording there, we saw only 
two young boys going up to the rooms (they 
never saw us). It seems unlikely that the current 
level of visitation will lead to degradation of 
deposits, though it will continue to wear away at 
the structures. Further public education 
concerning the fragility of these resources and 
more frequent passes by rangers going to and 
from Ceremonial Cave might help reduce this 
source of erosion. Perhaps the discreet 
placement of a rain shelter somewhere in the 
area would cut emergency visits by walkers 
caught far from the Visitor Center in summer 
storms. 

Natural deterioration at this group takes 
several forms. The most widespread results 
from the very friable nature of the lowest tuff 
unit in several parts of the Group (around A-l- 
A-9, A-23-A-29, and A-50-A-58, for example). 
This tuff erodes much faster than does the 
overlying stratum, and it forms the back wall to 
many rooms in the first visible story, as well as 
a possible, speculative story below that. The 
fact that it undercuts also threatens features in 
upper tuff levels, though this does not seem to 
be an imminent threat. Recent slumping of 
large tuff blocks has occurred in several places, 
most notably at A-72, parts of which seem to 
have fallen quite recently, and in front of A-15. 
In several places erosion and cliff deterioration 
seem liable to cause loss of features; this is true 
of the high rooms A-20 and A-39, where doors 
are very exposed and are becoming quite thin. 
A more severe loss could occur in rooms A-22 
and perhaps A-18, which are connected, 
complete rooms. The front wall of A-22 is 
supported by a very fragile-looking pillar, and 
loss of much of the front wall is conceivable. 
This condition might be relieved by repair or 
replacement of the masonry wall that appears to 
have been present there prehistorically. This 



CONTEXT 29 



course would involve at least partial excavation 
on both sides of the wall. 

Finally, there are three rooms with 
masonry closing portions of their fronts: A-10, 
A- 13, and A-60, all of which were pointed by 
Lister in 1939. The small doorjamb in A-13 
looks just as it did in 1939, and A-60 seems to 
be in reasonably good shape (Figures 2.7, 2.8). 
A-10 appears to be something of a miracle: a 
large piece of masonry appears to be supported 
mostly by fit and pressure; most of the wall has 
no foundation and forms a sort of arch (Figure 
2.9). Even more remarkable is the location of 
the room at the back of a large pour-off, which 
looks as though it should have washed the whole 
thing away long ago. An area in the middle of 
the wall now lacks mortar; at least pointing, and 
perhaps detailed recording, seems to be indicated 
here. The rooms along the side of the A-10 
drainage cut (A-9, A-ll, A-12) are severely 
eroded. 

In summary, the distinguishing 
characteristics of Group A are its proximity to 
the stream, its length, the presence of much 
rubble, an associated round masonry kiva in the 
flat area in front of the group, and highly 
variable tuff strata. 

Frijoles Group F, LA 50975 
(Held no. LA 50023) 

Of the groups recorded in 1986, Group 
F is central; it lies between Tyuonyi~of which it 
has a splendid overview~and the "Big Kiva." It 
also seems to have the most varied topography 
of the groups we studied, since it includes two 
large tent rocks and covers a considerable range 
in elevation (Figures 2.10, 2.11). The portion 
in which we worked may be considered the 
lower part in two senses-it is at the downstream 
end of the group and it is closer to the canyon 
floor. The area studied extends from a small, 
naturally sculpted arch at the upper end (this 
may well have been a room but was not 
recorded as such because it lacks features) along 
a stretch of relatively low cliff base behind two 



large tent rocks to just up the slope in the large 
embayment containing Group G. The tent rocks 
have remnants of rooms around their bases, and 
these rooms were included in the group both by 
Chapman and by our crew. The rooms from 
which data were collected can be contained in a 
trapezoid with a base of about 48 m along the 
cliff base, a height of 17 m from the cliff to the 
front of the tent rocks, and a top of about 20 m 
across the front of the tent rocks. The upper 
part of Group F is located on higher cliff bases 
and steeper slopes; we counted 46 rooms in this 
part of the group. 1 We recorded 48 rooms along 
the cliff base and 12 rooms around the tent 
rocks, so that the total number of cliff-associated 
rooms for Group F is 106. In counting rooms 
up- and down-canyon from specific rooms, the 
tent rock rooms were not included. Remnants of 
intact masonry exist, especially in the 
unrecorded part of the group; considering the 
centrality of this group both to Tyuonyi and to 
tourism, the rooms in this group seem to be in 
surprisingly good shape. 

Below where it levels out (at about F- 
15), the part of Group F that we studied has a 
great deal of masonry rubble associated with it. 
This segment of the group appears to be better 
suited to building than the upper part because of 
less slope. Approximately in the middle of our 
study group (rooms F-19-F-30), there may have 
been as many as six stories against the cliff 
(Figure 2.11), and considerable rubble remains 
in this area. Four cliff levels are clearly visible 
here, with a fifth suggested by a depression in 
the cliff above the uppermost. In addition, the 
viga holes for the lowest visible story are close 
to the present ground surface, which is the top 
of a considerable mound, leaving the possibility 
of yet another, invisible story at the base. This 
area shows evidence of a considerable 
expenditure of human energy, in the form of 
room remodelings, hand-and-toe hold trails, and 
large petroglyphs. There appear to have been 
retaining walls made of large blocks between the 
tent rocks. Terraces may also have been 
constructed at the extreme downstream end of 
the group, where the talus becomes quite steep 



30 CAVATE STRUCTURES 




Figure 2.7. Comparison photographs for A-l 3, showing little change during 47 years, a, b. Before- 
and after-stabilization photographs taken by Lister in 1939. c. Photograph of A-l 3 taken 
in August 1986. Note the similarity of plaster and masonry condition; somewhat greater 
wear is probably present at the lower left of the opening, and there may be additional 
graffiti on the plaster of A-l 4 below the lower right of the opening. 



CONTEXT 31 






Figure 2. 8. Comparison photographs for A- 
60. a, b. Before- and after- 
stabilimtion photographs taken 
by Lister in 1939. c. 
Photograph of A-60 taken in 
April 1987. Very little change 
is apparent after 48 years. 



32 CAVATE STRUCTURES 




Figure 2.9. Photograph ofA-10 taken in 1986. Lister did stabilization work on A-10 but did not 
include a photo in his report. It is remarkable that the masonry of this room, located at 
the head of a drainage, has survived. 



A 50975 

OUP F PLAN VIEW 



CONTEXT 33 



Boulder 
&.A- Bedrock 

• Rock Terrace 



5 

i i i i 




10 15 


20 2 

1 l 





10 
1 


meters 

20 30 40 

i i i 


50 

l 




1 sample. Shows Group F divided into upper and lower 
the rooms in lower Group F. Upper Group F has four 
bers shown for Lower Group F are those used in this 
'. in 1990 by J. Snead. 



CONTEXT 33 



LA 50975 

GROUP F PLAN VIEW 




Figure 2. 10. Room plan view for the Group F sample. Shows Group F divided into upper and lower 
parts; this project recorded only the rooms in lower Group F. Upper Group F has four 
sets of room numbers; the numbers shown for Lower Group F are those used in this 
study. Redrawn with an alidade in 1990 by J. Snead. 



CONTEXT 35 



ROFILE 



5 |inge (view point break) 

o 




Room not recorded by this study 
L — Rooms 19 and 20 obscured 
by angle in cliff face 

GROUP F 



sample. Shows Group F divided into upper and lower 
le rooms in lower Group F. Changes in cliff angle are 
s. Tent rocks with their cavate rooms are not shown. 
ndH. Newman. 



LA 50975 

GROUP F PROFILE 



CONTEXT 35 



Viga Holes 

Angle Change {view point break) 

5 10 15 




Room not recorded by this study - 
Rooms 19 and 20 obscured 
by angle in cliff face 



LOWER GROUP F 



Figure 2.11. Elevation sketch for the Group F sample. Shows Group F divided into upper and lower 
parts; this project recorded only the rooms in lower Group F. Changes in cliff angle are 
indicated by vertical dashed lines. Tent rocks with their cavate rooms are not shown. 
Redrawn in 1990, by J. Snead and H. Newman. 



CONTEXT 37 



as the cliff turns a corner into the recessed area 
containing the last few rooms of Group F (F-39- 
F-46) and then Group G. As at Groups A, I, 
and M, the distribution of rooms along the cliff 
base would have been very nearly continuous, 
though numbers of levels would have varied 
considerably. Especially above our study area, 
this distribution was accomplished in some 
places by heroic (if incomprehensible) room 
placements in drainages, in pour-offs, and on 
steeply inclined cliff bases. 

The elevations recorded for Group F 
rooms are based on an altimeter reading of 6160 
ft (1878 m) at the center of the sample. The 
quadrangle. in this area suggests 6180-6200 ft 
(1884-1890 m); the 6200 and 6300 foot contour 
lines are merged in this area. The 6160 ft (1878 
m) reading has been used as a baseline with 
some elevation added for the ends of the study 
group and some subtracted for the tent rock 
rooms. The elevations, once again, are 
approximate. The elevation of the Rito opposite 
Group F is about 6080 ft (1853 m). 

To my knowledge there are no records 
of excavation in Group F. Lister (1940a) did 
some stabilization in several Group F cavates, 
which we rephotographed (Figures 2.12, 2.13). 
Several of the walls he worked on have since 
collapsed. The proximity of these rooms to 
Tyuonyi and their ease of access makes it quite 
likely that they were among those used by field 
school students and earlier explorers, such as 
Bandelier. Indeed, in the vicinity of F-20, a 
row of nails has been driven into the cliff, 
presumably for clothes hooks. The old tourist 
trail passes below the Group F rooms, but not 
far below them. 

Frijoles Group I, LA 50974 
(Held no. LA 50022) 

Located at the top of a relatively steep 
and high talus, Group I is a compact group of 
rooms fronted by a reasonably level area around 
10 m wide (Figure 2.3, 2.14, 2.15). The main 



group of contiguous rooms is 29 m long (rooms 
I-7-I-35). At the upper end of the group are six 
rooms located above a precipitous drainage (1-1- 
1-6). Most of these rooms are lower than the 
main group and face more directly east. The 
lower end of the group is defined by a rockfall 
consisting mostly of very large boulders. Six 
rooms were partially obscured by part of this 
rockfall (I-30-I-35), and it seems likely that the 
event occurred after the construction, and 
probably after the abandonment, of the rooms. 
In addition, three rooms stand almost exactly 
midway between Groups I and J. It is not 
entirely clear whether Hewett and Chapman 
included the sole complete one (1-36) on their 
map, and since they are slightly closer to Group 
I, we included them in our version of Group I. 
(The intact room, 1-36, is the lowest of the three 
and is 18.2 m from 1-35 and 16.8 m from the 
highest Group J chamber). By our count, then, 
38 rooms can be seen on the cliff in Group I. 
All the rooms were recorded except 1-15, a 
small, third-level chamber we were unable to 
reach. Other rooms may exist at the same level 
as 1-15, but our inspection of some of these 
possibilities showed them to have no definable 
features and to be too ambiguous to be recorded 
as rooms. 

The flat area in front of the center of 
Group I has a considerable amount of rubble on 
it. There may have been two rows of masonry 
rooms in front of the cliff rooms. In the area of 
1-24, 1-25, and 1-27, the cliff rooms are three 
levels high, with the bottom level substantially 
filled. These filled rooms may have well- 
preserved floor features. 

By altimeter the elevation of the center 
of Group I is 6240-6250 ft (1902-1905 m). The 
map location is once again confusing. 
Converging contours are shown below the 
symbol that corresponds to Group I. This 
indicates a sharper, higher drop than is the case 
and suggests an elevation of around 6320 ft 
(1926 m). We used a baseline of 6250 ft (1905 
m) for the records; with the exception of I- 1-1-4 



38 CAVATE STRUCTURES 







Figure 2.12. Comparison photographs for F-31. a, b. Before- and after-stabilization photo- 
graphs taken by Lister in 1939. c. Photograph of F-31 taken in July 1986. There 
has clearly been a dramatic change since Lister stabilized this feature. In 
addition to the collapse of the entire masonry fissure closure, the viga hole at 
each figure's shoulder is broken away in the 1986 photograph, suggesting that 
vandalism may be involved. Note the lintel stone in the precollapse photographs 
and the groove left after its removal. 



CONTEXT 39 






Figure 2.13. Upper Group F, Room 12, a room not recorded by this 
project, as it appeared in 1939 and 1986. a, b. Before- 
and after-stabilization photographs taken by Lister in 
1939. c. Photograph of the same cavate taken in July 
1986. Note that the masonry plug to the right of the 
door is now gone and that the mortar has returned to its 
prestabilization state. 



CONTEXT 41 



50974 

IP I PLAN VIEW 

<- 1st Story Cavate 
L 2nd Story Cavate 

Wall Alignment 

Extent of Rubble 

Bed Rock 

Dripline 

Boulder 



5 10 15 

J_J I i_ 



40 SO 



rt all rooms are shown on the plan: some appear only 
, 21, and 29 are back walls and not shown). Redrawn 
lead and A. Prieto. 



CONTEXT 41 



Approximate Cliff Face 




Petroglyph 



LA 50974 

GROUP I PLAN VIEW 

-J ^ 1st Story Cavafe 

--' '. 2nd Story Cavafe 

oo cr ° Wall Alignment 

/n $s Extent of Rubble 

.£}£<■ Bed Rock 

Dripline 

£3 Boulder 



-LIMIT OF CERAMIC SAMPLE AREA 



Figure 2.14. Room plan view for Group I. Not all rooms are shown on the plan: some appear only 
in the elevation (Rooms 11, 16, 18, 21, and 29 are back walls and not shown). Redrawn 
using an alidade in 1990 by G. Head and A. Prieto. 



CONTEXT 43 



, Trails , 





Behind Rockfall 



LA 50974 

GROUP I PROFILE 

C I Boulders 
oooo Wall Alignment 
Viga Holes 



• ••• 



5 

_L_L 



10 



15 



20 



meters 

10 20 30 40 50 60 

i i I l I I I I 



feet 



drawn in 1990 by G. Head and A. Prieto. 



CONTEXT 43 



Petroglyph 



Hand- and -Toe Holds 

.15 



Cavates 30-35 Behind Rockfall 



LA 50974 

GROUP I PROFILE 



Trails 




Q 


Boulders 








OOOO 


Wall Alignment 








• ••• 


Viga Holes 










I I I I 


5 10 

I I I 




15 

I 


21 

I 




meters 










I 1 


10 20 30 

i i i 


40 

i 


50 
i 


60 

I 



feet 



Figure 2.15. Elevation sketch for Group I. Redrawn in 1990 by G. Head and A. Prieto. 



CONTEXT 45 



and a few third-level rooms, the rooms in Group 
I are at about the same level. The elevation of 
the Rito de los Frijoles below Group I is about 
6100 ft (1859 m). 

Trash is present on the slopes below the 
rooms, but as at Group M the quantity is fairly 
small. The length and steepness of the slope 
must again be considered, in addition to the 
possibilities of disposal elsewhere or in the 
rooms. Few artifacts are to be found on the 
surface of the masonry rooms or in the cavate 
rooms. 

Known previous work in Group I is 
limited to some stabilization done by Lister in 
1939 (1940a). Group I has several extant blocks 
of masonry, and he replaced mortar in some of 
these (including 1-10 and 1-22; Figure 2.16). 
Lister's photos of 1-22 show that a ventlike hole 
was completely blocked in 1939 but is now only 
about half blocked (Figure 2.16). The Civilian 
Conservation Corps (CCC) trail passes 
considerably below the group. The climb to 
Group I seems to be arduous enough to 
discourage visitors, so that recent impacts are 
much less than at Group A. The graffiti here 
are about as numerous as at Group M or perhaps 
slightly more so. Deterioration in central Group 
I seems not to be severe, but the ends have 
suffered more damage. The upper end in 
particular, which is located above steep slopes 
and presumably lost the protection of masonry 
fronts early, is severely weathered. Although 
some damage must have resulted from the 
rockfall at the lower end of the contiguous 
rooms, burial by rockfall may have had the net 
effect of preserving the lower parts of these 
rooms. 

There are two sets of hand-and-toe holds 
within Group I suggesting routes to the canyon 
rim. The first of these begins at around the 
middle of the group (above 1-27) and probably 
relied on rooftops as a starting place (Figure 
2.15). The second route is in the drainage in 
which the three isolated rooms (1-36-1-38) are 
located. There is a set of holds on each side of 



the rooms, leading to a bench and apparently on 
to the rim. Neither route was empirically tested. 
Other cliff features include sizable petroglyph 
panels at both ends of the contiguous cavates, a 
high petroglyph above 1-4, and one above 1-35. 



Frijoles Group M, LA 50972 
(Held no. LA 50020) 

Group M lies at the down-canyon 
extreme of concentrated sites of all types in 
Frijoles Canyon. The group is a long, fairly 
continuous set of cliff and masonry rooms above 
a sizable talus. The talus is neither so steep nor 
so high as that at Group I (Figure 2.4, 2.17, 
2.18). In terms of overall length and number of 
rooms, Group M is probably the largest group in 
which we worked in 1986, though Group A is in 
many ways comparable. Kohler estimates 
Group M to have had more than 200 rooms 
between a.d. 1325 and 1375 (Kohler and Linse 
1993:5), but Group M surface ceramics analyzed 
by the Bandelier Survey indicate that substantial 
occupation before 1450 is unlikely (Robert 
Powers, personal communication, 1994). 
Rainbow House (LA 217), a 50-60 room pueblo 
dating to the fifteenth century (see Table 2.2 at 
the end of this chapter; Caywood 1966), is 
directly below Group M and clearly visible from 
it. Some relationship is very likely to have 
existed between the two. Even closer to Group 
M is Saltbush Pueblo (LA 4997), an 11 room 
structure with a single pit structure. In terms of 
both architecture and ceramics, Saltbush Pueblo 
fits well into the Coalition Period, and the few 
absolute dates obtained fall in the thirteenth 
century (Snow 1974). David Snow suggests that 
occupation and construction of Group M took 
place during the latter stages of occupation of 
Saltbush Pueblo. His report compares the 
relatively low diversity of the faunal assemblage 
from Saltbush Pueblo (11 species) to the 
somewhat richer (14 species), very late Group 
M assemblage from Hendron's excavation, and 
to the very rich (23 species) Rainbow House 
assemblage. These three assemblages have little 
or no temporal overlap. 



46 CAVATE STRUCTURES 




Figure 2. 16. Comparison photographs for 1-22. a. After-stabilization photo taken by Lister in 1939. 
b, c. Two views of the door to 1-22 and the opening to 1-20, taken in 1986. Note again 
the lintel groove in the door; apparently the vent to the left of the door was completely 
closed in 1939, but most of the masonry is now missing. Much of the 1939 mortar is now 
gone, but the masonry above the door remains. The room may have had more fill in 
1939 than at present; other change appears to be minimal. 




CONTEXT 47 





O 












2 










0> 


• 










n 


in 










n 


3 










3 


O 










cc 


X 










o 

c 


o 

c 


» 

5 
55 


>» 
5 
35 


>> 
g 

0) 


5 
55 


£ 


£ 


■o 


■o 


£ 


X 


x 




c 






UJ 


UJ 


" 


CM 


CO 


» 


$ 


/ 


j 


* 




1 


J 


1 


f 


( 




( 


1 


1 


\ 






'> 



1 




I. 



.5 



5 






,60 



CONTEXT 49 




20 Linear meters of 
cliff face ommited here; 
elevation change is to 



s/l PROFILE 



ew point break) 
15 20 



25 



40 

i 



50 

i 



60 

_l 



sample. Upper Group M with vertical dashed lines 
Vooms 10, 61, and 65 are back chambers not visible 
\roup. Possible rooms (not recorded) are indicated 
1991 by K. Barthuli and S. Hall. 



CONTEXT 49 




20 Linear meters of 
cliff face ommited here; 
elevation change is to scale. 



Angle change, 

no cliff face omitted 



LA 50972 

UPPER GROUP M PROFILE 

i'""\ Possible Room 
^ Niche 
«•• Viga Holes 

Angle Change (view point break) 



10 
I 



15 



20 

I 



25 

_l 



10 20 30 40 



50 



60 

I 



feet 



Figure 2. 18. Elevation sketch for the Group M sample. Upper Group M with vertical dashed lines 
showing changes in cliff angle. Rooms 10. 61, and 65 are back chambers not visible 
when looking at the front of the group. Possible rooms (not recorded) are indicated 
by dotted outlines. Redrawn in 1991 by K. Barthuli and S. Hall. 



CONTEXT 51 



The cavate group is well defined by 
large drainage cuts at each end. The portion of 
the group in which we worked (upper Group M) 
extends from the up-canyon end, where there are 
two isolated rooms amid a rockfall, to about 
midway in the group, where a natural break is 
formed by some small tent rocks, a steep talus, 
and a slight lowering in the elevation of the 
cliff-base rooms. Upper Group M has an 
overall length of 102 m (78 m excluding the two 
uppermost rooms, M-l and M-2; Figure 2.17). 
Below the tent rocks multiple counts give an 
average of 73 rooms showing on the cliff. A 
total of 66 rooms out of 67 were recorded in the 
upper area, giving a total of 140 cliff-associated 
rooms in the group. Some high pockets in 
upper Group M have openings that look 
suspiciously like doors; those we were able to 
check were not rooms in our estimation, but 
some others may be. Recording of Group M in 
1990 by the Bandelier Survey confirmed the 
count of 67 rooms in the upper area, but the 
survey crew saw an additional 24 rooms (for a 
total of 94) in the lower area. Although it seems 
likely that their count includes features or 
pockets we did not consider rooms, the count 
discrepancy also reflects the intensity of cliff 
modification in lower Group M, and the 
potential this maze of features affords for 
deriving different counts. 

As at Group I, the width of the flat areas 
at the top of the talus varies. Toward the upper 
end (M-3-M-14) the talus is steep and hard and 
extends directly to the base of the cliff; in this 
area no evidence of masonry rooms remains, 
though they were unquestionably present. In 
front of M-18-M-20 the top of the talus is 
considerably flatter, and there are many 
suggested walls and strewn blocks. Here it is 
first possible to see three levels of rooms (M-17 
and M-64). In front of M-25-M-31 the amount 
of rubble further increases, and in this area three 
levels of rooms are again visible on the cliff. 
M-33, once an elaborately decorated room, 
stands on the up-canyon edge of the largest 
concentration of rubble for the entire group, 
including the part below our study area. Here a 



slight embayment in the cliff contains Rooms M- 
35-M-49, and the rubble extends 9-10 m out 
from the cliff base with a relief of around 2 m. 
Wall alignments suggest as many as three rows 
of rooms out from the cliff, and there are as 
many as five levels of cavates (Figures 2.17, 
2.18). Given the height of the mound and the 
level of rooms such as M-54, a large structure is 
indicated. Several cavate rooms in this area are 
larger and very heavily used and have many 
features. This area is "downtown" for Group 
M. Past the mounded area, in front of Rooms 
M-50-M-60, there is once again a steep, 
disintegrated tuff talus slope, though more 
masonry blocks are present on this slope than at 
the upper end. Even in this location, cliff 
evidence shows rooms on two to three levels, 
most of which seem likely to have had masonry 
fronts. Some large pieces of canyon wall have 
fallen here, which probably removed natural 
fronts of some of these rooms. The rooms 
above these steep slopes seem to have suffered 
the most from natural disintegration, presumably 
because masonry elements in these locations 
were far more prone to collapse. 

Lower Group M contains several notable 
features. As discussed earlier, the best 
documented excavations of cavate rooms in the 
canyon are Hendron's (1943) excavation and 
stabilization of four cavates and five associated 
masonry rooms (see also Turney 1948 and 
Kohler's [Carlson and Kohler 1989] excavation 
of a single cavate chamber nearby). There are 
several exposed hearths in Lower M, two of 
which were sampled in a pilot archaeomagnetic 
study. At the bottom end of the group is a large 
chamber located above a major drainage; the 
chamber was probably subdivided during at least 
part of its use. This "kiva" contains rock art, 
including a relatively well-preserved, green 
Awanyu. 

Almost all the rock art at this group is 
confined to room walls. Room M-33 contains 
remnants of red, yellow, black, and white wall 
paintings, as well as incised figures, on at least 
two walls and on several coats of plaster. Its 



52 CAVATE STRUCTURES 



condition is now rather fragmentary and 
warrants careful removal and expert study of 
what remains. 

A route leads out of the canyon not far 
up-canyon from the group, but none seems to be 
located inside the room cluster. There is a set 
of six large holes above the four-level section of 
upper Group M (above M-41), but these seem to 
lead only to a ledge where possible room M-64 
is located (Ranger Ed Greene reached this area 
by rope). 

The cavates and features we recorded in 
Group M are on the same level to a remarkable 
degree. We did not use an altimeter for this 
group, but the contours here are reasonably 
clear, and the 6180-6200 foot (1884-1890 m) 
level seems to identify the location of the 
majority of the rooms (in several places there 
are multiple levels of rooms). Using 6200 ft 
(1890 m) as a baseline for the main level of 
rooms (M-14, M-18, M-21, M-23, M-33, M-39, 
M-43, M-53, M-56, and-though they are well 
above the base of the cliff-M-59 and M-60), we 
estimated other elevations. Most other rooms 
are within 5 m of this elevation. The general 
level of upper Group M is around 6 m higher 
than the lower half. Group M is not indicated 
by a symbol on the USGS Frijoles Quadrangle, 
but Group L is suggested and shows as 6300 feet 
(1920 m) because of merged contours. This is 
probably too high; the elevation of the Rito 
below Group M is about 6040 ft (1840 m). At 
Rainbow House the Rito enters a narrow and 
fairly deep inner canyon, making it much less 
accessible to irrigation than in the main part of 
the canyon. 

Tsankawi Section, LA 50976 
(Held no. LA 50024) 

Cavate rooms are abundant in many of 
the dry canyons north of Frijoles Canyon. With 
a few exceptions, such as Puye and Garcia 
Canyon, these features have received less 
attention from archaeologists than those in 
Frijoles. Lister (1940b) did some stabilization 



and mapping in several areas that were later 
divested from the monument, as well as at 
Tsankawi. Lister's work resulted in several 
maps, some of which cover long stretches of 
cliff and great numbers of cavates, as at Otowi 
and Tsankawi. 

This area takes its name from Tsankawi 
Pueblo (LA 211), a major "plaza pueblo" 
located on the mesa top with a commanding 
view in all directions. The mesa (or potrero) is 
formed by Sandia Canyon on the south and Los 
Alamos Canyon on the north. Tsankawi is a 
Tewa word meaning 4 the place [or gap] of the 
round cactus'. The orthography of Native 
American languages has been modified many 
times. Thus, a more recent, presumably more 
nearly correct, rendering of Tsankawi is 
Tsankawi'i, and earlier ones include Sankawi 
and Sankewi. The NPS and USGS use 
Tsankawi, and I have used that spelling for ihe 
ruin and as shorthand for the cavate group we 
studied, although there are cavates around 
Tsankawi Pueblo not included in this study. To 
make matters even more complicated, there is a 
pottery type named after LA 211; since it was 
named early, it has traditionally been spelled 
Sankawi Black-on-cream. The pottery name, 
too, has been rendered in many ways, usually 
without the initial tee (e.g., Sankawi'i [Kohler 
and Linse 1993:36]); the traditional spelling is 
used here. 

Cavates are abundant on the south and 
east slopes of the mesa and can be separated into 
reasonably discrete groups on the basis of 
distribution, as was done long ago for the 
Frijoles cavates. Such groups include the rooms 
located at the end of the western lobe of 
Tsankawi Mesa (not recorded by Lister), the 
lower rooms in the red tuff (Lister's C-123-C- 
165 and C-170-C-174), the rooms in the south- 
facing rincon below the pueblo to the southwest 
(Lister's C-69-C-120), and the east-facing rincon 
below the pueblo to the southeast (Lister's C-15- 
C-45 and C-50-C-67). Such subdivision may 
impose some breaks that do not correspond to 
prehistoric groupings, and other methods could 



CONTEXT 53 



be used to distinguish other groups. Still, this 
division has the advantage of creating more 
manageable analytical units for archaeologists. 

The east-facing rincon group was chosen 
for recording because it contains many rooms 
(even more than we anticipated) in both the red 
and gray tuff units. Also, the rooms in the red 
unit looked to be in good condition relative to 
most of the other rooms in that stratum. The 
"site," as we defined it, is delimited by two 
points in the mesa side. The point on the north 
and east extreme is small, and there are scattered 
cavates quite close to the LA 50976 group in the 
next rincon to the east. The point on the south 
is a more distinct break formed by a relatively 
high, sheer cliff and an absence of cavates for a 
considerable distance. 

Lister mapped and mentioned 49 rooms 
in this area. With a few exceptions, he included 
only relatively complete chambers and did not 
assign separate numbers to back chambers. Our 
count for the whole area is 119 rooms, reached 
by counting exterior rooms indicated by cliff 
features and by giving back chambers separate 
numbers. We retained Lister's numbers but 
replaced his C prefix with a TS (C indicated 
only "cave" and was used at all the other groups 
Lister worked in as well). New TS numbers, 
beginning with 501, were assigned to the rooms 
he had not numbered. TS-501 was chosen as a 
starting point because Lister's numbers for the 
whole Tsankawi area go up to 180, and it was 
thought desirable to flag the added numbers and 
keep them separate (TS-501-TS-570 were used). 
Lister's map clearly shows the relative locations 
of structures, and we adopted it as a location 
map to which we added our new numbers. The 
map included here (Figure 2.19) is based on the 
modified Lister map and a detailed sketch map 
drawn by the Bandelier Survey crew. The only 
other recorded work in this group is the removal 
of a secondary burial by Johnson (1960; see 
chapter 1 of this study). 

The rooms in the LA 50976 rincon are 
located on about four different levels, three in 



the thick gray-and-white tuff layer, and another 
in the underlying red tuff layer (Figure 2.20). 
Within the rincon several subgroups can be 
defined; these could be considered as the 
equivalent of room blocks in a masonry site or 
as separate sites. At the north end of the rincon, 
on the highest bench, is a group of about 25 
rooms (TS-15-TS-26 and TS-501-TS-513) 
located in a porous, frothy tuff layer, overlain 
by the homogeneous, obviously indurated layer 
that forms the mesa caprock. They are closely 
spaced, and most front on a flat shelf 4-5 m 
wide that may have had structures on it: there 
are several sets of viga holes for rooms near 
rooms TS-21-TS-25, even though masonry 
rubble is scarce. Above most of this stretch of 
rooms on the rimrock is an elaborate, varied, 
and well-preserved rock art panel; the good 
condition of this panel results from the hardness 
of the top stratum. Among the petroglyphs there 
are three large rectilinear depressions ground 
into the cliff face. These features seem to occur 
only here and with a second cavate group 
recorded by Lister in a rincon to the west of 
LA 50976. After some discussion, we decided 
to call these features cliff niches, though their 
bases slope somewhat (Figure 4.23). There is 
also a cleavage in the rimrock with paired beam 
seats reminiscent of scaffold supports, like those 
at Scaffold House in Navajo National 
Monument. The densest cluster of rooms is on 
the next level down, on the east-facing slope and 
about in the center of LA 50976. Rooms TS-54- 
TS-61 and TS-530-TS-533 are packed onto 
several levels, and rooms TS-62-TS-65 are close 
by. Down yet another level is a smaller group 
of rooms, the last cluster in the white and gray 
tuff: TS-51-TS-53 and TS-540-TS-546. 

The top of the red tuff layer forms a 
distinct ledge for most of the length of the 
rincon. This coarse, very soft tuff, possibly the 
Tsankawi Pumice Bed (Bailey et al. 1969:14- 
15), is exposed for thicknesses of 3 m or less for 
most of the middle of the rincon, with thicker 
exposures at each end. The rooms in the red 
tuff form a subgroup because of their elevation; 



Petroglyphs 




Petroglyphs 



Petroglyph 



Red Tuff 



*^?Pf-i \ 



f ** 43 ^ \«A 

565 564 



CONTEXT 55 



LA 50976 

PLAN VIEW 



D 


Petroglyph 





Bedrock Trail 


^^ 


Bedrock Cliff or Outcrop 


— . — 


Ceramic Sample Areas 




Oripline 


^ 


Talus Boulder 


5 

i , . , i i 


10 15 20 




meters 




i 


10 20 30 40 50 

i i i i i 




feet 




Scale Approximate 



25 



506 
507 



\ y, Petroglyphs 



-501 



Hand -and -To* Holds 

/ 



| Limit of Northern Ceramic Sample Unit 

, ..': Hand -and -To* Holds 



<le. First mapped by Lister (1940b), locations for this 
py of Lister's 1940 map; remapped by R. Powers and 



CONTEXT 55 




LA 50976 

PLAN VIEW 



O Petroglyph 

Bedrock Trail 

'j &V&H Bedrock Cliff or Outcrop 
— — Ceramic Sample Areas 

Dripline 

Talus Boulder 

10 15 20 25 



10 20 30 40 50 

i 1 i i i i 

feet 
Scale Approximate 



Hand -and Toe Holds 
I 

| Limit ot Northern Ceramic Sample Unil 

... Hand and Toe Holds 

C »ProclT 



Figure 2. 19. Plan view for the Tsankawi sample. First mapped by Lister (1940b), locations for this 
study originally recorded on a copy of Lister's 1940 map; remapped by R. Powers and 
T. Chadderdon in 1988. 



CONTEXT 57 



LA 50976 
PROFILE 




wi sample. Given the curvature and stepped nature of 
necessarily somewhat schematic; relative positions and 
proximate. This drawing is based on field sketches and 
R. Powers in 1994. 



CONTEXT 57 



LA 50976 
PROFILE 




42 562 

Hand and To« Holds' 



557 

Hand and To* Hold* 



Figure 2.20. Elevation sketch for the Tsankawi sample. Given the curvature and stepped nature of 
these cavates, this drawing is necessarily somewhat schematic; relative positions and 
shapes are shown, but scale is approximate. This drawing is based on field sketches and 
photos; redrawn by S. Herr and R. Powers in 1994. 



CONTEXT 59 



from TS-29 to TS-39 the rooms are fairly 
continuous, including TS-548-TS-558, TS-567, 
and TS-568. The southernmost rooms form a 
final red tuff subgroup: TS-40-TS-45 and TS- 
559-TS-566. The cliff exposure here is greater, 
and these rooms are built on several levels. 
Some isolated rooms are interspersed among 
these groups, and some pockets are also found, 
which may have been modified for storage cists 
but were not recorded. The pockets are 
especially abundant in the white tuff at the south 
end of the rincon where no rooms are present. 
Elevations were assigned to the various levels 
beginning at 6600 ft (2012 m) for the uppermost 
level and dropping by 20 foot contours to 6500 
ft (1981 m) for the red tuff rooms. To judge 
from Lister's photos, there has been little change 
in the condition of cavates at Tsankawi over the 
last several decades (Figure 2.21). 

Because of the multilevel, dispersed 
nature of LA 50976, trails and staircases are 
abundant. Lister shows six stairways, and worn 
trails connect most of the groups. Some of these 
may be the result of the heavy visitation this 
area has had and continues to receive~at the 
very least, modern visitors surely enhance 
prehistoric trails. Other paths, however, now 
terminate where there would once have been a 
roof (e.g., in front of TS-507). None of these 
routes was separately recorded. 

A primary reason for recording some 
structures at Tsankawi was to look for degrees 
of difference that might speak (however subtly) 
to the question of differences in ethnicity 
between the inhabitants of Frijoles and 
Tsankawi. Anthropological tradition holds that 
Frijoles Canyon is the northern limit of 
prehistoric Keres speakers, with the Tewa 
extending to the north (see, for example, Hewett 
1938; Steen 1977). Accordingly, about a third 
of the 1986 field time was spent at Tsankawi, 
and about the same proportion of records was 
compiled there (8 full field days and 119 
records). Differences are present among all the 
groups in which we worked, but the differences 



are greatest in the case of Tsankawi. This site 
decidedly stands apart from the Frijoles groups. 

Some of the differences between Frijoles 
and Tsankawi relate directly to topography. In 
contrast to the sheer cliffs of Frijoles Canyon, 
the sides of Sandia Canyon below Tsankawi 
Ruin consist of short cliffs separated by benches 
and talus slopes. This affects both the 
distribution of rooms and construction 
possibilities: the dispersed, many-tiered 
distribution of rooms at LA 50976 is just not 
possible in Frijoles, and the absence of high 
cliffs at LA 50976 means that multiple stories 
would have been far more difficult to construct. 
Where cliffs are higher at Tsankawi, such as 
around TS-35-TS-39, there is evidence for 
multistory structures. Another noticeable 
difference between Frijoles groups and 
LA 50976 is in exposure. While the Frijoles 
groups were pleasantly cool and shady (the 
unappreciative said cold) until late morning in 
July and August, the LA 50976 rooms received 
early-morning sun, with some late afternoon 
shade. 

More likely to have cultural (or 
temporal?) significance are several differences in 
the presence and absence of features, and in 
their characteristics. Rooms at Tsankawi are 
notably larger than those in Frijoles, both in 
height and floor area. While at least two 
examples of two chambers being joined into a 
single larger chamber were found at LA 50976, 
none was identified in the larger Frijoles Canyon 
sample. Remains of masonry structures, 
masonry plugs, and small walls are scarce at 
Tsankawi. It is possible that the cavate rooms at 
LA 50976 were more often primary rooms while 
those in Frijoles were more often back chambers 
of secondary importance, though the features in 
many of the Frijoles chambers are not in accord 
with this interpretation. Room outlines, 
particularly at Frijoles Group A, suggest that 
large chambers were in use in Frijoles as well. 
The scant evidence for masonry at LA 50976 
may be related to the presence of postholes both 
inside and outside the rooms there, while none 



60 CAVATE STRUCTURES 




■ 






■ ■ J- 


MH 


^^^v% " MM 










"» 





Figure 2.21. 



Comparison photographs for TS-53. a, b. Before- 
and after-stabilization photographs taken by Lister in 
1939. c. TS-53 as it appeared in August 1986. 
Other than in vegetation, little change is apparent. 



CONTEXT 61 



was identified in Frijoles; however, Frijoles has 
few of the expanses of clear horizontal tuff that 
are common at Tsankawi. The proportion of 
nearly complete chambers is much higher at 
Tsankawi than at the Frijoles sites. Excavated 
small back chambers are also relatively more 
abundant at LA 50976. 

The extramural rock art at Tsankawi is 
in a different class from what is now visible in 
Frijoles, in terms of quantity, variety, and 
preservation. It is, of course, difficult to 
determine how far preservation is responsible for 
the other differences. We observed several 
features only at Tsankawi: the cliff niches 
mentioned above, vertical holes in ceilings 
(named Panowski Holes in honor of one of our 
crew), groups of floor pits, and series of deep 
incisions in room walls. We observed other 
features in Frijoles that seem to be lacking at 
LA 50976: slots, metate rests, and floor ridges. 
Loom support-and-anchor complexes seem to be 
especially common at both Group M and 
Tsankawi/LA 50976. 

"Cuevitas Arribas" (no LA number) 

In addition to the formal recording of the 
sites described above, our crew also conducted 
a preliminary reconnaissance of a group in 
Frijoles Canyon not recorded on the Hewett- 
Chapman map. This group is located upstream 
from Ceremonial Cave and at approximately the 
same level on the canyon wall (Figure 2.1). 
Unlike those farther down the canyon, it is 
located on a bench above the first cliff above the 
creek. Although the canyon is narrower here, 
the placement of this group is probably related 
to the presence of a low bench by the Rito and 
a wider floodplain, which could have provided 
an opportunity for irrigated farming similar to 
that at Group A. There are three subgroups 
separated by drainages. The rooms closest to 
Ceremonial Cave are in very poor condition. 
There are no complete chambers, and the degree 
of exposure and the coarseness of the tuff here 
has left few recognizable features. The upper 



two parts, located just past the wilderness 
boundary, are closer together, larger, and better 
preserved. We named this site Cuevitas Arribas 
(loosely, "the cavates above"), to indicate its 
location up-canyon from Group A. Separate 
room counts came to 72 and 74 rooms in the 
two upper subgroups; including the 10-15 less 
well-preserved rooms down-canyon, this 
settlement contained well over 80 rooms. There 
are several examples of well-preserved masonry, 
including a retaining wall inside a chamber, a 
dividing wall, some very clear room outlines in 
front of the cliff, and some partial closures of 
openings. One of the rooms has the reddest 
plaster of any observed during this study. In 
spite of its exclusion by Hewett and Chapman, 
the Cuevitas Arribas group is equivalent to 
Groups A-M and definitely merits further study. 

Cavate Chronology 

Attempts to decipher relationships 
among cavate groups, between cavates and large 
masonry structures, and between purported 
Tewa and Keres areas all hinge on tight 
chronological control. Sadly, few dates are 
available from cavates, and this project adds 
little to the broad outlines already known. P. J. 
McKenna and R. P. Powers conducted a field 
analysis of the surface pottery at the groups we 
studied, but the sherds are generally not 
abundant and all are found on the talus slopes 
below the cavates. The location of the ceramics 
means that their association with structures is 
vague and there is high potential for mixing. 
With a very few exceptions, dates can be 
assigned only to groups of cavates and not to 
individual features. McKenna's report on the 
surface ceramics is given in full at the end of 
this chapter. 

Incorporation of Ceramic Dates and 
Further Dating Potential 

Based on McKenna's analysis we 
assigned the following date spans (all dates are 
a.d.) to the groups we studied: 



62 CAVATE STRUCTURES 



Group A 


1225-1550 




Group F 


1100-1175 


1315-1550 


Group I 


1275-1525 




Group M 


1250-1550 


1695-1725 


Tsankawi 


1300-1650 





Archaeomagnetic sampling of two exposed 
hearths in lower Group M (behind Hendron's 
excavations and between there and M-60, the 
lowermost cavate we recorded in the group) 
provided dates of a.d. 1425-1550 (±26) and 
1275-1550 (±41) (Table 2.1). Together with 
the Group M tree-ring dates of 1493 and 1494 
(if the dates given by Smiley et al. [1953] and 
Robinson et al. [1972] are in fact different 
specimens), occupation of Group M seems quite 
firmly placed in the 1400s. The archaeo- 
magnetic dates suggest that the extent of the 
eighteenth-century reoccupation was small, since 
the hearths are very near where the late 
materials were excavated. The small quantity of 
late ceramics further corroborates that 
suggestion. 

Justin Hyland's (1986:50-54) ceramic 
analysis indicates that the Garcia Canyon cavates 
he studied fall toward the early end of the dates 
derived for the Frijoles and Tsankawi cavates. 
He suggests a range of 1275-1350 and a mean of 
1275, though his analyses were not complete 
when he wrote his thesis. None of the tree-ring 
samples submitted yielded dates (Hyland 
1986:86). The decorated type dominating cavate 
assemblages analyzed by Hyland is Santa Fe 
Black-on-white, to which he assigns a date of 
1275 in calculating his ceramic dates (Hyland 
1986:61, appendix). 



the south; in addition to Hyland's ceramic counts 
at Garcia Canyon, Stuart and Gauthier 
(1981:105) place the beginning of Puye at ca. 
1250. Major differences in the period of use of 
cavates are unlikely, however, given the 
relatively small area of the Pajarito Plateau. All 
the areas included in the present study are near 
large, aggregated pueblos that postdate the 
manufacture of Santa Fe Black-on-white, and 
both the ceramics and the few chronometric 
dates suggest primary use during the fourteenth 
and especially fifteenth centuries. Santa Fe 
Black-on-white, however, is present in all the 
counts in small quantities. Given intensive use 
of existing cavates and the probable construction 
of more rooms during aggregation, it is easy to 
understand how Santa Fe Black-on-white would 
become a minority type, especially in surface 
collections. 2 

A plausible sequence is that cavate 
structures came into use relatively early in the 
intensive occupation of the Pajarito Plateau, 
around 1200. As the population aggregated, 
they continued in use and were a part of the 
aggregation phenomenon. Given their 

permanence relative to masonry structures, some 
cavates were periodically reused into the historic 
period. Like other forms of habitation on the 
Pajarito, then, cavates were constructed in the 
Coalition Period and continued in use in the 
Classic Period, though the settlement pattern 
changed rather dramatically. Because many 
cavates remain intact and unfilled, the total use 
span of a cavate that survives intact is potentially 
very long-much longer than for almost any 
masonry room not continuously maintained. 



The high frequency of Santa Fe Black- 
on-white on these sites is in accord with 
Gauthier's assessment that many cavates, 
especially more dispersed ones, were constructed 
and used before the establishment of large 
aggregated sites (R. Gauthier, personal 
communication, 1989). There is also some 
indication (again based nearly entirely on 
ceramics) that cavates may have been in use 
earlier in the northern part of the plateau than in 



The long potential use spans of unfilled 
cavate rooms complicate precise dating of 
construction and period of greatest use. As 
always, the best hope for precise dates lies with 
tree-ring samples. Though William Robinson 
and others (1972:73) indicate that more samples 
from cavates may be in collections, including 
some from Group F, the only two dates now in 
the record as having come from cavates are 
those from Group M (Table 2.1), and there is a 



CONTEXT 63 



Table 2.1. Chronometric Dates and Ceramic Associations from Bandelier. 



Site 



A. Adapted from Smiley et al. 1953:21-39 



Samples 



Dates 



Ceramics Associated 



Group M 


1 dendro 

2 archeomagnetic 


1493c 

1275-1550±41 

1425-1550±26 


Biscuit A and B 

Sankawi B/c 

Rio Grande Glazes I-V 


Tyuonyi 


16 dendro 


1383+-1466c 
1417-1505 best 


Kwahe'e B/w, Santa Fe B/w 
Wiyo B/w, Biscuit A & B, 
Sankawi B/c 
Rio Grande Glaze I-V 


Rainbow House 


20 dendro 


1421-1453 
1448-1451 best 


Biscuit A & B, Sankawi B/c, Santa 
Fe B/w, Rio Grande Glaze I-V 


Frijolito 


3 dendro 


1431-1447 


Kwahe'e B/w, Santa Fe B/w, Wiyo 
B/w, Biscuit A & B, Rio Grande 
Glazes I-III, White Mountain 
Redware 



B. From Robinson et al. 1972 



Site 


Samples 


Dates 


Comments 


Puye 


41 dendro 


1413-1562w 
1536-1577r 


miscellaneous and unknown 
proveniences; no cavates specified 


Tschirege 


33 dendro 


1411-1581v, w 
1559r, 1572 + 


miscellaneous proveniences; no 
cavates specified 


Tyuonyi 


55 dendro 


1386-1467r 
1309-1527v, w 


construction 14th- 16th centuries; 
Group F nearby 


Group M 


1 dendro 


1494vG 


Hendron excavation; see part A 
above 


Saltbush Pueblo 


3 dendro 


1194-1241w 


see Snow 1974; near Group M 


Big Kiva 


15 dendro 


1505-1525r 
1383-1525v, w 


below Group I 


Note: All dendro dates 


are ranges for terminal dates only (outside rinj 


*s only, pith not given). 



64 CAVATE STRUCTURES 



chance that they come from the same specimen. 
Although some original structural wood probably 
exists in unfilled cavates, we observed none 
during this project. Datable wood from cavates 
is more likely to come from excavation of at 
least partially filled rooms and possibly hearths. 

Other sources of date ranges include 
further archaeomagnetic sampling, obsidian 
hydration analysis of artifacts from firm contexts 
(no suitable artifacts were recovered during this 
nonexcavation project), and C-14 dating. The 
heavy carbon deposits on the ceilings of many 
cavates open the possibility that C-14 samples 
may be available without excavation. Hyland 
(1986) submitted some such samples, though he 
reports no results. These samples pose even 
greater problems than most C-14 samples: the 
source of the carbon may be material that died 
long before it was burned, the carbon present is 
likely to have come from numerous different 
burning events, and there is no control over 
what the original substance was. Given the 
inherent imprecision and ambiguity of C-14 
dates, their high processing cost, the relatively 
fine ceramic dates possible, and the additional 
problems with these samples, they are of little 
potential utility, though still tantalizing. 

Analysis of Surface Ceramics from the 
Study Areas 

Peter J. McKenna 

In recording surface ceramics at the five 
recorded cavate groups, we followed procedures 
adopted in the 1985 pilot survey at Bandelier 
(McKenna and Powers 1986). Given the 
absence of other means of dating the cavates, 
ceramic dating was the focus of the data 
collection, but we also recorded vessel form. 
Here we present the cavate chronology indicated 
by the surface ceramics and point out some 
differences among ceramic assemblages of the 
cavate groups as suggested by surface materials. 



Sampling Areas 

Cavate refuse scatters are diffuse. In 
covering approximately 32,400 m 2 of sample 
area in two days, we located and identified only 
2553 sherds. This gives an average of 0.08 
sherds per square meter, in stark contrast to the 
extremely high-density scatters around Tsankawi 
and Yapashi Pueblos proper, where surface 
sherd scatters were as high as 254 sherds per 
square meter and no lower than 87. In the 
sample survey, of the 21 fully recorded sites 
only a trail and a small structure showed 
ceramic densities lower than those recorded at 
the cavates (McKenna and Powers 1986: Table 
8). We noted no dense ceramic trash in cavates 
or along the flat rubble/talus margins 
immediately fronting cavates, but we did see 
some variability in the relative frequency of 
ceramics among cavate scatters. To locate 
ceramics we scoured extensive areas of rubble 
mound or talus slope in front of recorded cavate 
groups. We collected, identified, and tabulated 
the ceramics, and replaced them by scattering 
them in the same general area from which they 
were collected. We found no ceramics in any 
Frijoles cavates during ceramic sampling and 
observed only four Biscuit B sherds in separate 
cavates at Tsankawi. We observed a few sherds 
in two inaccessible Frijoles cavates (F-56 and 
M-15) during cavate recording, but none was 
included in this analysis. We sampled Groups 
A, F, I, and M in Frijoles Canyon; we did not 
sample the "Cuevitas Arribas" site. We 
subdivided the cavates at Tsankawi into three 
sampling units and those at Frijoles Group F 
into two. 

All sherds observed in the following 
areas were included in the counts: 

Group A: the talus/rubble fronting the 
west half of the group, from the central kiva to 
cavate A-l (108 m E-W x 20 m N-S, ceramic 
n=253). In addition two 2 x 1 m grids were 



CONTEXT 65 



sampled at 16 m and 34 m east of the central 
kiva in the soil-fan bordering the talus/house 
rubble. 

Group F: the talus/rubble mound 
fronting the entire group of architecturally 
sampled cavates. Two cavate groups were 
apparent: those constructed on the south side of 
some detached tent rocks and those along the 
main cliff at the top of the talus. The lower 
group was sampled in an area (24 m N-S by 23 
m E-W, ceramic n=263) directly downslope, 
while the remaining area was associated with the 
upper cavates (ca. 41 m E-W x 30 m N-S, 
ceramic n=282; total ceramics n=545). 

Group I: rubble mound and talus 
fronting the architecturally sampled cavates (60 
m N-S x 30 m E-W, ceramic n= 174), to near 
the base of the talus slope. 

Group M: rubble mound and talus 
fronting cavates, primarily between drainages 
flanking the central portion of the architecturally 
sampled cavates, erosion having denuded slopes 
along the east and west margins of this cavate 
group (55 m E-W x 40 m N-S, ceramic n=672). 

Tsankawi Group (LA 50976): three 
sample areas, a northern and central group 
located in gray and white tuff just below the 
mesa top, and a lower or southern group in the 
red tuff (see Tsankawi description above). The 
northern group consisted of cavates below the 
east-west trending section of hard caprock (an 
irregular 58 m E-W x 30 m N-S, ceramic 
n=362). The central group consisted of a 
scatter focussing on the middle group of cavates 
in the white tuff (50 m NE-SW x 45 m NW-SE, 
ceramic n=141), with the lower margin of 
collection ending at the top of the lower red tuff 
layer. The southern, or lower, sample area 
fronted cavates below this red layer of tuff (91 
m NE-SW x 17 m NW-SE, ceramic n=406). 
The south cavate sample was taken in halves for 
comparison because the northeast half of this 
sample area lay below the cavates in the upper 
gray and white tuff and undoubtedly contained 



some ceramics from this upper scatter. The 
ceramic sample from the Tsankawi LA 50976 
cavates is 906 items. 

Chronology and the Rio Grande Series 

Ceramic chronologies and type 
descriptions are based on information presented 
in the sample survey report for Bandelier 
(McKenna and Powers 1986) and summarized in 
Table 2.2. Unidentifiable or chronologically 
nonsensitive categories are often composites of 
technically separate classes. An example is 
"whiteware," which is undecorated portions of 
bichrome service wares, whether those 
specimens are technically black-on- white or 
black-on-cream. Cavate sites at Bandelier have 
seen heavy visitation, and for decades collectors 
have relentlessly gathered their ceramics. This 
has resulted both in selective disappearance, 
particularly of decorated sherds, and in sherds of 
a generally smaller, nondescript character- 
important points in the evaluation of cavate 
ceramics. 

Ceramics and the chronological data 
base in the Rio Grande region have, at best, a 
working relationship between two independent 
data sets that requires constant scrutiny and 
evaluation. The default time spans assumed 
(Table 2.2) in dating the present samples are, 
naturally, only as accurate as the patchy 
information on which they are based. A healthy 
degree of skepticism and periodic reevaluation of 
ceramic chronologies are necessary because such 
constructs rest on the broad extension of limited 
chronometrics, with some inherent imprecision, 
and on the necessity of accepting uniformity 
across space as a characteristic of the "type" 
ceramics. The extremely short time spans 
assigned to types such as San Lazaro Glaze-on- 
polychrome are suspect and should be used only 
tentatively (F. H. Ellis, personal communication, 
1985). While there is no particular reason to 
suspect that ceramic chronology on the Pajarito 
Plateau is totally out of step with the Rio Grande 
chronological sequence, some differences are 
likely (see Lang 1982). The chronological data 



66 CAVATE STRUCTURES 



Table 2.2. Date Ranges Used by Various Analysts for Rio Grande Ceramic Types. 





McKenna & 


Hubbell & Traylor 


Snow 1982f 


Breternitz 


Ceramic Type 


Powers 1986 


1982:242 


Warren 1979 


1966 a 


Kwahe'e B/w 


1000-1225 


950-1225 


950-1225 


1125-1200 


Socorro B/w 


1100-1250 


1050-1275 




1050-1275 


Santa Fe B/w 


1175-1300+ 


1175-1300+ 


1175-1300 


1200-1350 


Wiyo B/w 


1250-1400 


1250-1400 


1300-1400 


1300-1400 


Biscuit A 


1350-1450 


1350-1450 


1350-1450 


1375-1450 


Biscuit B 


1425-1550 


1425-1550 


1425-1550 


1400-1550 


Galisteo B/w 


1300-1400 


1250-1400 


1250-1350 


1300-1400 


Sankawi B/c 


1500-1675 


1500-1675 


1550-16251 


1500-1600 


Tewa Polychrome 


1675-1720 


1675-1720 


1675-1720 


post 1500 


Kapo Black 


1650+ 


1650+ 


1650+ 




Posugue Red 


1675 + 




1675 + 




Potsuwi'i Incised 


1450-1550 


1450-1550 


1450-1550 


1425-1525 + 


Glaze A 


1315-1425 




1300-1450t 




Agua Fria G/r 




1315-1425 


1315-1425 




Cieneguilla G/y 




1325-1425 


1325-1425 




San Clemente G/Poly 




1315-1425 


1315-1425 




Cieneguilla G/Poly 




1325-1425 


1315-1425 




Glaze B 


1400-1450 


1400-1450 


1400-14501 




Largo G/y-r 






1400-1450 




Largo G/Poly 










Glaze C 






1450-1490t 




Espinosa G/Poly 


1425-1490 


1425-1490 


1425-1490 




Glaze D 






1490-15 15t 




San Lazaro G/Poly 


1490-1515 


1490-1515 


1490-1515 




Glaze E 






1515-1625f 




Puaray G/Poly 


1515-1650 


1515-1600 


1515-1650 




Pecos G/Poly 


1515-1700 


1515-1700 







CONTEXT 67 



Table 2.2. (continued) 



Ceramic Type 



McKenna & 
Powers 1986 



Hubbell & Traylor 
1982:242 



Snow 1982f 
Warren 1979 



Breternitz 
1966 a 



Escondido G/Poly 


1515 + 


1515 + 




Encierro G/Poly 








Glaze F 






1625-1700t 


Kotyiti G/y-r 






1650-1700+ 


Kotyiti G/Poly 









a Base reference Smiley et al. 1953. 



in use at present are unstructured and inadequate 
beyond a very crude "regional" level. 

The chronology of the sequence itself 
rests on widely scattered, poorly provenienced 
data (Smiley 1951; Robinson et al. 1972), 
summarized and ceramically correlated by 
Smiley and others (1953), and taken verbatim in 
the most "recent" gospel on ceramic dating for 
Rio Grande types (Breternitz 1966). Dates 
available from Bandelier sites poorly reflect the 
temporal range of ceramics actually present 
(Table 2.1). Certainly more recent work, with 
better provenience/ceramic associations, is 
available for a reappraisal of the general Rio 
Grande sequence and perhaps for a more 
rigorous scrutiny of areas within the region 
(Orcutt 1994). 

Variation from accepted temporal 
sequences by subregional ceramic assemblages 
has been noted by Mills (1986) along the lower 
Rio Grande and is just as likely on the Pajarito 
Plateau. Differences in date assignments to the 
same pottery types by projects in the Rio Grande 
region are another indicator of subregional 
ceramic temporal variation (see Table 2.2). 
Both more extensive dating of Pajaritan contexts 
and a complete summary of more recent work 



are needed before a Pajaritan chronology can be 
redefined and compared to the general Rio 
Grande sequence. Ceramic chronologists 
traditionally track "trade" items on a separate 
time line than indigenous wares because nonlocal 
types tend to occur longer in regions removed 
from the production center. Use of the default 
time spans for these trade types, which are 
assumed to be a substantial component in 
Bandelier ceramic assemblages, further muddles 
the assignment of site chronology. 

Cavate Group Ceramic Dates 

The ceramic samples from the Bandelier 
cavate groups suggest variation in initial 
occupation and periodicity in reuse. All cavates 
show occupation during the Coalition and 
Classic Periods in the fourteenth, fifteenth, and 
early sixteenth centuries; only one, Frijoles 
Group M, shows early eighteenth-century 
ceramics (Table 2.3). In general, based on 
relative frequencies of ceramics, early 
occupations are most evident in the upper 
Frijoles groups, while later occupations are more 
evident in cavates farther down-canyon. 
Tsankawi cavates are largely contemporaneous 
with the mesa-top pueblo a.d. 1250-1650), but 
differences in the relative abundance of Sankawi 



68 CAVATE STRUCTURES 



1 

a 



o 



1 

1 1 
< 1 



* 



ll 



a 55 
e2 



PL, 



«£ 



1-2 

.8 £ 



«? 



|55 
H 

CO 



U 






2 
o 

U 





t- 




c> 


© 






u") 




m 


-N 




m 




ON 


■«t 


»n 


On 


«-H 






U"> 




d 




>— i 


en 






— i 




e*> 


d 




T* 




^H 


d 


1-H 


^* 


d 






CN 




VO 




2 


o 






CN 

— 




00 

o 

en 


- 


00 

5 






VO 


en 


CN 


VO 


- 






*— < 

CN 


PH 


•* 


VO 


o\ 


i-H 


ts 


CN 


^ 


<n 


ON 


t- 




r» 


vo 


oo 


cn 


CN 


C-; 




^H 


en 


Tf 


d 


d 


oi 


od 


m 


d 


d 


*"" ' 


d 


ir> 


d 






d 


VO 


d 


d 


en 




d 


d 


CN 


— < 


r» 


5 


</■> 


»n 
>0 


en 


en 


o\ 


Os 


00 

on 


CN 


5 

»— c 




vo 


vo 


en 


<* 


CN 

VO 




i—i 


m 


§ 


^^ 


CN 


00 


en 


8 


en 




Ti- 


v© 


1 


en 


ON 

5? 


00 

On 




w-> 


— 


CM 


VO 
CN 






en 


>-H 




(S 


en 


■<t 


<s 








- 


00 
ON 




o 


CN 

en 
vo 




■* 






C~ 






(H 


f-H 




en 


■<t 


t- 


CO 




en 


en 


<N 


00 

en 


CS 


Ti- 
en 


o 

VO 


VO 




'- 




CN 
CM 




- 


- 1 


00 






On 


•■' 


en 

CM 






CM 




00 


t- 


o 

CN 
CN 


cn 

e'- 




00 


""* 


CN 


f- 








o 

*-H 




CN 


v© 


t- 


■<* 








NO 


00 






en 




cn 


■<t 


vo 


vo 


oo 




00 


00 




d 


rj 


d 


^H 
















CN 




d 


d 


f— 4 


m° 


On 




r»' 


ON 




■-H 




r~- 


(•I 








on 

On 


§ 




o\ 

<* 






CN 


tj- 


m 


en 


en 

ON 






00 



CN 


■* 


vo 


*n 


es 


(S 


C^ 




ON 


d 


^^ 


•" ' 


Tt 


d 




d 




VO 


cn 


en 


<ri 


tj- 


es 


ON 


ON 


ON 
ON 




cn 


>TN 


en 


en 




o 


VO 
en 


VO 
CM 


q 

NO 




en 


■^■ 


en 


CM 


oo 
en 


CM 


00 


NO 

ON 




i/"> 


00 


in 




vo 

00 


en 




en 

0\' 

vo 



q »h »n Tt on r~; en 

'H d vo vd ■-! en -h 



ON —I 



ON 


00 


r~ 




en 

o 


CO 


en 


r-~ 


en 


5 



-H -H »-< U-) VO l/^ 



en o rN i/"> >tn t* 



1 1 

S3 03 
00 00 



2 1? 



3 



S> "S ^ "8 

c 



a -c 



"1 E 1 

a o a 

M V &o 



5 



CO 



U Oh 



"S -8 

1 



T3 

' 1 



<u 



a 



I 

Oh 



.s 

a 

2 "^ 
» J 

a -c 

Oh 00 



11 
a « 

U CL, 






« o i < a .*: 5 U S3 

^oo>0QOaooOHOHl> 



CONTEXT 69 



a 



o 



* 








00 






2 o 






VO 








8? 


oo 


-H 




o 


~H CS 


ro 




6 


6 




ON 


6 d 


6 


1* 


1-H 


^H 


CS 

•— 1 




^ -<t 


>/-> 



Uh 



£ 



K 



3 £ 

CO 



H 
o 

2 
3 



VO —l 



CO 

© 



•i & 
r-8 



■t o 



m 



oo 



cs 

CO 



00 

cs 



co oo 



CS On 
CO 



On CO 

ON ^ 

cs 



1 

3 1 

5 o 

Z % 

« I 

§ 1 

a ft 



VO CS 

^ 6 



CO cs 



tJ- vo vo 

6 6 © 



m ifl m 



6 



VO 

6 



Ifl H H N -H ^-, 

dodo do 



^h co co 



t- o 



0\ - h tn 



"<t -H 



CS cs 



CO 

d 



w> 



o 
cs 



CO 



NO 


ON 


ON 

d 






CS 


00 

cs 


On 

CS 

On 




CS 


cs 




On 


NO 


CO 


d 




4 


d 


d 


cs 


co 




VO 


CO 



CO VO 

d 



cs 



CO 



-H CS 



On 



On 



CO ~h 

d d 



CO ~* 



d 



CO 



d 



CO t-» 



1 -s 5 

« « £ 

« .2 o 

^ a, Oh 



3 U 



o 



>> 




•3 J 




►, 8. o 




t-i ^ *^r 




3 S 




*< 5. & JJ OQ 


o o o 


Agua 
Ciene 
Ciene 
San C 
Glaze 


o o o ^ 


00 60 00 4J 
H 1- «- S 







j*. 


>» 




a 


1 




o 


o 




o 


i 

o 


Q 


1 


.a 


m 


J 


& 


d 


§ 


u 


n 


CO 



70 CAVATE STRUCTURES 



5 






6 



o 



a 
i 



* 



3 o 



1 * 

o 
H 



ou 



* 



la a 



* 



H 

CO 



U 



o 



© d 



m ^ , 



**! *i °. 

00 vo On 



»-H ^H Tf 

m w> r- 



o\ ^ « ^ 
oi d « d 



<S no -H 



N * 0\ 



^ h m 



on 









(S 



d do d 



w 



O 



3 lfl|ll 

& I § I i ~ | 

^ u fc o O O P 



en 



oo 

o 



oo 






oo 






»o On 



o o 



"> oo 



o 

On 
NO 



(S 00 






ON ON 

m • 

fN 



<n o 
o 



» m ^ 



c- 

d 



3 «> 



J3 u 



3 

"3 § 






own 

CSV© 



vnvriON(S 

ts>nNor^ 



V/"Nl/-> 

Nifl 



8u">w>© 



VTNO 







<s 


t"» 




ON 




1250- 
1650- 
or 1675 


>n 


^ 


c» 




■<1- 




o 




*-< 


d 


d 




rJ 




8 




















- 1 


NO 






s 

On 






■* 




Tf 


ON 


c4 


8 


3 


fONO 


en 




CS 


T* 


oo 


5 


«n 


1 

do 










rJ 


vi 


©u-> 

PHNO 
















*-* *-H 


NO 


- 1 


- 


ON 




NO 

m 


00 

ON 
CO 


1 

©>/-> 

CO NO 



3 



CONTEXT 71 



Black-on-cream suggest either that cavates were 
not occupied as long or that the use of Sankawi 
Black-on-cream differs at the cavates. 

In the Frijoles groups, Group A (1225- 
1550) is estimated to have been first occupied at 
1225 because of the amount of Santa Fe Black- 
on-white relative to Wiyo and later 
black-on-whites. Group F (1100-1175 and 
1315-1550) seems to have seen two separate 
occupations, one during the late Developmental 
Period and another during the Classic Period; 
the absence of Santa Fe Black-on-white suggests 
that Group F was not in use during the Coalition 
Period. This pattern is consistent in both cliff 
cavate and tent rock ceramic subgroups at Group 
F. With only two Santa Fe Black-on-white 
vessels represented, Group I (a.d. 1275-1525) 
shows little evidence of a late Coalition Period 
occupation; at the terminus of its ceramic 
sequence, Glaze D polychromes are more 
similar in rim form to earlier Glaze C forms 
than to the Glaze D examples encountered in 
other cavate groups. These facts suggest that 
Group I may have had the shortest period of 
occupation of all the sampled cavate groups, 
with a late beginning date and an earlier ending 
date. Group I also has the fewest ceramics 
tallied, though the sample area is comparable to 
those covered at other cavate groups. This low 
frequency of sherds may also be a reflection of 
a shorter occupation span; whatever the cause, 
a sample size problem is evident which is critical 
in assessing the validity of the temporal 
placement of the group. Group M (1250-1550 
and post-1695 [to 1725]) shows the only clear 
evidence of early historic ceramics. A single 
Kapo Black and historic polychrome jar base 
were the only representatives of this late 
assemblage in the 1986 sample. Turney's 1948 
work (see Table 1.1) presents other evidence for 
Pueblo Revolt-early eighteenth-century 
occupation at Group M. The mid-1200s- 
mid- 1500s occupation at Group M is indicated 
by the continuum of decorated types from Santa 
Fe through Biscuit B and Glaze E ceramics. 
Glaze D seems to have lasted longer than the 
traditional 20-year span, ending in 1515, while 



the Glaze E sample is largely finished in muted 
fawn and tan tones, a style of coloration likely 
coeval with late Glaze D and predating what 
may be the later use of light cream slips in 
Glaze E. This is the basis for the terminal date 
assignment of 1550 for the initial occupation of 
Group M. No Glaze F was identified at any of 
the cavate groups sampled. 

Ceramic samples from the Tsankawi 
Pueblo proper suggested a date of 1250-1675 
based on conventional time spans for the 
production of Sankawi Black-on-cream in 
conjunction with minor but consistent amounts 
of Santa Fe and Wiyo Black-on-white. 
Beginning dates nearer 1300 and ending dates 
nearer 1650, however, may be more 
appropriate. Cavate samples suggest a similar 
time span for those features. No confidently 
identifiable glaze types were found in the cavate 
samples, but the impression from partial rims 
and surface attributes of glaze sherds is that 
glaze import became more common during and 
after Glaze D; a similar situation was noted in 
the mesa-top pueblo samples as well. 

Two other cavate groups in Frijoles 
Canyon, sampled during the 1985 test survey, 
also had ceramics indicative of late occupation. 
Cavates B-70 and B-71 (portions of Groups C 
and B respectively as defined in the 1985 pilot 
survey [McKenna and Powers 1986]) had even 
fewer sherds than the present sample of cavates, 
but they had early historic types of Kapo Black, 
Tewa Polychrome, and Glaze E. These cavate 
groups were dated 1450-1700 and 1500-1700 
respectively. Here again, sample size problems 
undoubtedly affect the assessment of occupation 
span. Cavate pueblos sampled away from the 
main clusters in Frijoles Canyon and around 
Tsankawi show earlier, more discrete occupation 
spans. LA 50909, south of Frijoles Canyon in 
the Corral Hill area of Bandelier National 
Monument, was primarily a Coalition Period site 
dating to 1200-1425 (McKenna and Powers 
1986: Table 9). Coalition Period occupation 
was also the main period for cavates surveyed 
north of Bandelier by the Pajarito Archaeological 



72 CAVATE STRUCTURES 



Research Project (Hill and Trierweiler 1986). 
Especially when viewing sherds on the surface, 
the long-term aggregation associated with 
communal pueblos and cavates is a major factor 
in making occupational episodes at cavate groups 
indistinct. The trend to aggregation is stronger 
at lower elevations (such as Tsankawi) than at 
higher ones (such as at Yapashi). 

Ceramic Patterns 

The chi-square statistic was used to test 
for differences in the ceramic samples from the 
cavates. Vessel forms, reduced to open 
("bowls") or closed ("jars"), ware, and 
provenience, were the dimensions of the tests. 
Data and test results are presented in Table 2.4. 
The 0.05 level was accepted as significant. 

In the tests of intrasite sample areas for 
Group F and at Tsankawi, we found no 
differences in the general vessel populations. 
Clearly, surface ceramic evidence cannot be 
used to support special use of cavates at the 
intrasite level as presently measured. However, 
considering only decorated ware in the Tsankawi 
red tuff cavates (south sample area), there is a 
significant difference in vessel form between the 
cavates in the western and eastern halves of that 
subgroup (Test 1): a greater than expected 
number of whiteware jars is found in ceramics 
associated with the western red tuff cavates. 

In the comparisons between areas at 
Tsankawi (Test 2) we found a significant 
difference between the northern area and the 
central and southern areas. The number of jars 
is higher than expected in the northern area, 
while bowls are more common in the central and 
southern areas; northern and southern area 
distributions are the strongest contributors to this 
pattern. Distributions of service ware only 
(Test 3) reinforce this observation, with bowls 
being more common in the southern and central 
areas and jars more common in the northern 
cavates. A test of ware distributions (Test 4) 
shows culinary (exclusively jars) more than 
expected in the north area, which again 



contributes to the value of the chi-square in 
determining a significant difference in form 
distribution between these cavate subdivisions at 
Tsankawi. In comparing cavate ceramic 
distributions with those of the pueblo, we found 
general agreement in form ratios (bowls :jars- 
Tsankawi= 1:2.9; cavates = 1:2.9), but the 
number of decorated bowls (Test 5) is higher 
than expected at the cavates, as compared to the 
main pueblo midden sample, where jars are 
more common. 

We found no significant difference in 
form distributions among the Frijoles cavates. 
Form distribution, however, does seem to differ 
between cavate sites and some communal 
pueblos. A test (Test 6) of distributions of 
culinary versus service ware shows that culinary 
ware is significantly more abundant at cavate 
sites, while service wares are relatively more 
common at communal pueblos. Since culinary 
ware consists almost entirely of closed forms, it 
follows that closed forms will be relatively more 
common on cavate sites in later tests. 
Comparison of all forms among the Frijoles 
cavates, Yapashi, and Tsankawi Pueblo plus 
cavates (Test 7) shows the Yapashi and 
Tsankawi samples to be higher than expected in 
bowls while the Frijoles cavates are higher in 
jars. A test of the Tsankawi Pueblo and cavate 
ceramic forms showed no significant difference, 
permitting us to treat them as a unit in Test 7. 
A test considering only service wares (Test 8), 
however, shows that the Tsankawi sample is 
higher than expected in bowls and the Frijoles 
cavates are higher in jars, while forms at 
Yapashi occur as expected. Some differences 
may be related to trends in ware forms: glazed 
ware production is more often associated with 
closed forms and Tsankawi's matte-paint 
bichrome tradition with bowls. Some 
differences may also be related to differing 
emphasis in site functions involving ceramic use. 
However, the consistently low value for the 
coefficient of contingency (Table 2.4) indicates 
a weak association between site types and the 
ceramic dimensions of form and ware in the 
present data. 



CONTEXT 



73 



Table 2.4. Ceramic Form and Ware Data and Significance Tests. 





All Ceramic Forms 3 


Test No. 


Decorated Only 




Site 


Bowls 


Jars 


Bowls 


Jars 


Test No. 


Frijoles cavates 














A 


47 


206 




47 


26 




F 


119 


424 




119 


80 




I 


38 


136 




38 


26 




M 


110 


553 




110 


70 




B-70 (C) 


14 


56 




14 


15 




B-71 (B) 


21 


54 




21 


14 




Tsankawi cavates 
North 


69 


280 


2 


69 


29 


3 


Central 


40 


101 


2 


40 


17 


3 


South 


120 


285 


2 


120 


21 


3 


Frijoles cavates A-M 


314 


1319 


7 


314 


202 


8 


Tsankawi cavates 


229 


666 




229 


67 


5 


Tsankawi Pueblo 


255 


645 




255 


150 


5 


Tsankawi all 


484 


1311 


7 


484 


217 


5, 8 


Yapashi Pueblo 


234 


572 


7 


234 


124 


8 


Intra-site 














Group F 
Lower 


53 


210 




52 


43 




Remainder 


66 


214 




66 


37 




Tsankawi (south sample) 
East half 


83 


196 




83 


9 


1 


West half 


37 


89 




37 


12 


1 




All Ceramic Wares 










Site 


Culinary 


Decorated 


Test No. 


Tsankawi 
North 


251 


98 








4 


Central 


84 


57 








4 


South 


264 


141 








4 


Tsankawi cavates 


599 


296 








6 


Frijoles cavates 


1104 


540 








6 


Yapashi Pueblo 


448 


377 








6 


Tsankawi Pueblo 


495 


450 








6 



74 CAVATE STRUCTURES 



Table 2.4. (continued) 



Test No. 


N 


x 2 


degrees of 
freedom 


probability 


Contingency 
Coefficient 


Cells <5 


1 


141 


5.456 


1 


0.020 






2 


895 


10.250 


2 


0.006 


0.106 





3 


296 


9.216 


2 


0.010 


0.174 





4 


895 


7.929 


2 


0.019 


0.019 





5 


701 


16.596 


1 


0.000 






6 


4309 


84.469 


3 


0.000 


0.139 





7 


4234 


39.476 


2 


0.000 


0.096 





8 


1575 


8.834 


2 


0.012 


0.075 






"Unknown forms excluded. 



The distinction in ware varieties between 
Tsankawi and Frijoles needs no test to illuminate 
it. All samples show culinary ware as the 
majority of the ceramic assemblage. The 
culinary wares are distinct between the two 
areas. The micaceous paste Tesuque series 
dominates the Tsankawi assemblage, and the 
sand-tempered utility wares of the Rio Grande 
series are most abundant in the Frijoles cavate 
sites. The pattern in decorated or nonculinary 
ceramics is even more apparent: glaze-paint 
ware is clearly associated with the Frijoles 
cavates (23 percent of the assemblage), while 
nonglaze service wares are associated with 
Tsankawi (31 percent). Although the ordering 
of the ware representations is clear, there is 
relatively more matte-paint material in the 
Frijoles sample (9 percent) than there is 
glaze-paint material in the Tsankawi sample 
(2 percent); see Table 2.3. 

Summary 

As might be expected, the surface sherds 
inventoried at the Bandelier cavates provide date 
groups too broad to be useful for fine temporal 
comparison or structuring in analyses of 
architectural variability. All architecturally 



sampled cavate groups overlap for some 
undeterminable period(s) during their 
occupation; it is not possible to isolate "earlier" 
from "later" groups of cavates either within or 
between sites based on the ceramic sample. 
Temporal differences are present, but the nature 
of the ceramic samples does not permit their 
correlation to groups of cavates within sites or 
the meaningful temporal placement of cavate 
sites relative to one another. 

Temporal holes exist which may relate 
to some cavate variation, but confidence in 
assessing those "holes" depends on the 
willingness to accept chronological evidence 
from very sparse, widely scattered, and heavily 
picked-over ceramic assemblages. The present 
sample suggests the following major temporal 
distinctions: (1) the early, apparently 
noncontinuous, occupation of cavate Group F; 
(2) the post-Pueblo Revolt historic reoccupation 
of Group M; (3) the apparently relatively short, 
possibly less intense, occupation at Group I. 
These distinctions generally suggest that 
occupation of cavates was discontinuous, at least 
in the Frijoles groups. The spotty evidence of 
historic reoccupation is clear at Groups M, B, 
and C. Prehistoric periodicity in occupation is 



CONTEXT 75 



also suggested by stylistic trends in some cavate 
groups, such as the similarity of Glaze D to C 
rims in Group I, and by the apparent 
discontinuity in some type sequences (though 
this may be reading more into the sample/type 
tabulations than is warranted). At Tsankawi no 
temporal distinctions are evident among the 
cavates or between the cavates and the main 
pueblo. 

Some functional differences are 
suggested between cavates and between cavates 
and nearby communal pueblos. A difference in 
functional emphasis at cavate sites is implied by 
the greater-than-expected number of culinary 
vessels. The association of closed forms with 
the Frijoles cavates is particularly strong, since 
both culinary and service- ware jars occur more 
often than expected. Such a difference might 
indicate use of cavate sites for living and 
cooking, or it might result from an emphasis on 
storage in cavates. Better samples and more 
detailed analysis of excavated samples are 
needed to refine this interpretation. While 
culinary ware forms the majority at all sites 
examined here, we suspect that this may be at 
least in part a function of long collection of 
decorated sherds. Some differences in the 
proportion of service ware in sections of the red 
tuff cavates and in vessel forms between the 
cavate subgroups at Tsankawi also suggest some 
functional variation within cavate groups. 



Economic and possibly ethnic differences 
are certainly reflected in the relative proportions 
and types of wares in the Tsankawi and Frijoles 
samples. The relative proportions of micaceous 
to sand-tempered culinary, and matte-bichromes 
to glaze-paint ware, are reversed in the two 
areas. This clearly indicates differing patterns in 
ceramic circulation and, by extension, in spheres 
of production and exchange. Nevertheless, 
functional patterns of ceramic use, when 
compared with a limited sample of community 
pueblos, tend to crosscut these differences in 
both the Tsankawi and Frijoles cavate samples. 
This indicates a measurable difference in site 
types, which may be more clearly expressed in 
architecture than in ceramics. 



1. This count includes all traces of former rooms. Figure 
2.10 shows at least 32 excavated rooms. 

2. Santa Fe Black-on-white is also a minority type at 
Tyuonyi and, as at the cavates, its presence appears to 
reflect early occupation. Although Tyuonyi was 
probably constructed between the late 1300's and early 
1500's, tests conducted by Onstott (1948) below the 
earliest surface in Tyuonyi's plaza revealed Santa Fe 
Black-on-white and contemporary utility wares 
suggestive of a late Coalition occupation. A higher 
plaza surface with Santa Fe Black-on-white and early 
glazes is either associated with the remains of an early 
structure or the early building stages of the present 
pueblo (Onstott 1948; Van Zandt 1994). 



Recording Procedures, Group Attributes, 
and Cavate Condition 



Recording cavates presents an interesting 
archaeological problem. On the positive side, 
many features not normally visible without 
excavation are exposed without the time and 
effort required for digging. On the negative 
side, cavates are difficult to survey because they 
are numerous, because they are exposed to 
various types of degradation, and because 
recording all the data available would consume 
much field time. To take advantage of these 
data-rich, well-preserved features, a survey on 
the Pajarito Plateau must compromise between 
recording a few cavates in extreme detail and 
recording many of them in insufficient detail. A 
major goal of the 1986 fieldwork was to develop 
a procedure for efficiently recording enough 
cavate data to understand variability in their 
construction and features. Although it was not 
so used, the recording method was intended for 
dealing with cavates encountered by the survey. 
The sample of cavates and associated features 
recorded by this pilot study provides baseline 
information on cavate variability, as well as on 
the condition of this particular set of features. 

Recording Procedures 

Cavate and Noncavate Forms 

The prehistoric Pueblo people had many 
techniques for using the tuff cliffs of the Pajarito 



Plateau. In some cases they hollowed entire 
chambers out of the cliff, leaving a small door. 
In others they seem to have excavated a large 
chamber and then built the front wall of the 
room entirely of masonry. The proportion of 
room excavated into the cliff ranges 
continuously from a fully excavated chamber 
with a tuff front wall, to a masonry-fronted 
completely excavated chamber, to masonry-tuff 
combination rooms, to mostly masonry rooms 
with back walls somewhat indented, to rooms 
that used the otherwise unmodified cliff with 
some viga holes bored in it. Edgar Hewett 
recognized and described this range at Puye: 

We note here three classes of dwellings. 
1. Excavated, cavelike rooms, serving 
as domiciles without any construction in 
front. 2. Excavated rooms with open 
rooms or porches built on in front . . . 
3. Houses of stone, one to three stories 
high, with corresponding terraces, built 
upon the talus against the cliff. In these 
groups the excavated chambers now 
seen in the cliff wall were simply back 
rooms of the terraced buildings. . . . 
An examination of the talus discloses 
remains of several villages of 
considerable extent that were built 
against the cliff. (Hewett 1908:19) 



77 



78 CAVATE STRUCTURES 



This variability adds another dimension 
to the problem of recording sites with cavates. 
The situation is further complicated by processes 
of erosion and falling blocks of cliff, and by the 
fact that masonry is more susceptible to 
disintegration than are cavate rooms. To attempt 
to address the problem, we used two types of 
forms in the field, one for cavates and the other 
for noncavates. The noncavate form was 
designed for rapid recording of rooms showing 
little cliff excavation. Such rooms clearly lack 
many of the features that more completely 
excavated rooms may have (Figure 3.1). The 
difficulty comes when deciding which form to 
use for borderline features. Copies of the forms 
as they were used in the field, the coding 
conventions, and the dates of added categories 
are given in appendix 1 . 

Our use of these forms was influenced to 
some extent by the evolution in our perception 
of the project's goal. Initially the focus was on 
cavate structures as unusual architectural features 
in need of better recording, with a nod to other 
associated features. Early in the field season we 
adopted a somewhat different and, we believe, 
better approach: the recording of whole sites. In 
two respects, we did not attain the goal of full- 
site recording. First, we noted but did not 
record in detail architecture that had no 
manifestation on the cliff, such as rubble 
mounds and walls. Second, we did not record 
artifact samples. The latter shortcoming was 
partially remedied by a later field inventory of 
surface ceramics (see chapter 2). 

The use of two forms allowed us to 
separate features into analytical classes and also 
shortened the time spent recording partial 
features. Reduced recording of partial features 
is acceptable because they contain less aggregate 
information and waste many form entries. As 
our working definition of a cavate in this study, 
we specified that a feature must be enclosed on 
at least three sides and have some remaining 
sheltering qualities. Some judgment is required 
to gauge the completeness of a cavate room, and 



the decision is complicated in some cases by 
erosion of the exterior edge. We recorded 
cavates we considered reasonably complete and 
made an ordinal estimate of the completeness of 
all cavates. The noncavate category includes 
badly collapsed former cavate rooms, 
excavations in the tuff for rooms that were 
largely built rather than excavated, and other 
prehistoric, artificial cliff features. Completely 
natural pockets or overhangs showing prehistoric 
use should be recorded with their natural origin 
noted, but during our recording we found no 
examples of cavates we regarded as completely 
natural. 

In the field the distinction between a 
cavate and a noncavate usually centered on 
whether the feature was judged to be sufficiently 
intact and to contain sufficient information to 
merit detailed recording. Early in the recording 
(especially at Group M) there was greater 
variability in application of differentiation 
criteria. On the whole, as the season progressed 
we tended to make greater use of the noncavate 
form in the case of partial rooms. Group M has 
a relatively low number of noncavates, which is 
partly an artifact of this adjustment period and 
partly the result of the occupants of Groups F 
and A having made greater use of the cliff as a 
back wall with no excavation than did the 
builders of the other three groups. A 
standardized list of noncavate feature types did 
not exist during the field recording. The 
categories we used, however, fell into a fairly 
restricted range and were placed into the 
following groups during the data correction 
phase: back wall, partial cavate, filled cavate, 
chamber, hand-and-toe hold route, cliff niche 
(see also Table 3.1). The cavate form includes 
an estimation of room completeness, which thus 
becomes an important screen for chamber 
analyses, since it allows sorting of observations 
into reliability categories. 

Another carryover from the initial 
inclination to treat the two categories differently 
is that features within noncavate rooms were 



RECORDING PROCEDURES 79 




Figure 3.1. Schematic drawing of three cavates and two noncavates, showing feature types. At upper 
left is a row ofviga holes defining a noncavate backwall; at lower left is a mostly filled 
cavate that would have been recorded on a noncavate form in 1986. At mid-left and 
upper right are the most common form of cavate, which would have been closed by 
masonry, and at lower right is a fully enclosed cavate with an exterior door. Note the 
truncated pyramid shape, and how wall plaster extends only partway up the wall. 
Ceilings are smoke-blackened and show digging stick marks. The features shown, with 
code numbers, are as follows: 



2 Exterior door 
4 Firepit 

6 Floor ridge 

7 Floor pit 

8 Large floor-level niche 

9 Wall niche 



10 Slot 

11 Viga holes 

13 Loom anchors 

14 Upper loom supports 

15 Smokehole 

16 Vent 



19 Interior door 

26 Exterior opening 

27 Ceiling 

28 Indeterminate holes 
37 Wall depression 
40 Metate rest 



80 



CAVATE STRUCTURES 



sometimes measured by ranges (which was our 
initial intent) and sometimes individually (which 
we concluded was preferable). Features 
recorded only within a range obviously cannot 
be used the same way as individually recorded 
ones (though ranges are decidedly faster to 
measure and record). In the spirit of full-site 
recording and monitoring of the full range of 
variability, the two-form approach had its uses, 
but the forms should have been fully congruent 
in terms of codes and feature recording. It is 
probably preferable to have a single form with 
a rigorous set of room types, along the lines 
enumerated by Hewett (1908; see above). 

Although we had both computer and 
paper forms for recording data, only the paper 
form is presented here and in appendix 1 . The 
computer form was slightly different, but all the 
same information was recorded in both formats, 
and all data are, of course, now equivalent in the 
various data bases. Certain classes of 
information do not apply to certain sets of 
features (firepits do not lead to other rooms, and 
a wall does not need a precise location). These 
classes of features are separated on the paper 
form, with only the variables relevant to each 
included. We stressed the avoidance of 
redundancy in recording, so that notes relative to 
specific features, such as condition of walls, 
were entered with the features rather than with 
general room notes. 



portion was usually a straightforward 
assessment, though partitioning walls in curved 
chambers sometimes required subjective 
judgments. Azimuth location of features and 
measurement of height above floor give a 
relative location of all features relative to all 
others. The compass was usually located by 
establishing the intersection of the diagonals 
across the room. Taking an azimuth to a feature 
requires reducing it to a point; we shot to the 
center of features. Long Awanyu petroglyphs 
and whole walls are not amenable to this 
treatment, so we did not take azimuths to them. 
Ideally, every feature would also have a distance 
measured to the "Brunton station" (the compass 
location), but this procedure would greatly 
increase the recording time. As a compromise 
and a means of better locating walls, we adopted 
the practice of taking angle and distance to 
distinct chamber features. Chamber corners 
were given preference, but other features 
occasionally had to be used. Given these basic 
measurements it would be possible either to 
relocate the measurement point with considerable 
accuracy or to create a rough plan of a room 
from measurements. The measurements would 
be especially useful for reconstructing feature 
layout to record station-to-feature distances to all 
floor features, though we did not do so in 1986. 
Rooms that did not have enough space or floor 
for setting up the compass lack azimuth locations 
for features. 



Discussion of Codes 



Shapes 



Although all the "variable states" are 
presented in appendix 1 with the forms, several 
variables require further explication, and some 
need reconsideration after field use. 

Chamber Location 

Locations of features within cavates were 
recorded with a two-part system: by chamber 
portion, such as right or left wall or floor, and 
by means of azimuths from a point as near as 
possible to the center of the room. The chamber 



Given the variety of chamber and feature 
shapes present in the cavates, it is clear that 
measurements are far more meaningful if 
associated with a shape. This requires 
visualizing features as regular geometric figures, 
which is often not easy. The effort is 
worthwhile, however, because it makes the 
measurements more accurate and appropriate 
while also serving to describe the features. The 
use of fractions further refines the information 
collected. 



RECORDING PROCEDURES 



81 



Geometric Fraction 

The two fraction codes-geometric and 
feature—probably caused more discussion and 
confusion than any other part of the recording 
process. Geometric fraction is intended to be a 
descriptive code and to complement the 
assignment of a shape to a feature. Because 
features often do not conform well to purely 
geometric shapes, geometric fraction is an 
estimate of how much of the shape assigned and 
measured is actually present. For example, if a 
chamber most nearly approximates a rectangular 
solid, but a portion of the whole solid is missing 
because of curvature of the cliff at the opening 
of the room, estimating the amount of the solid 
that is present tells the analyst that the volume of 
the room is perhaps only 80 percent as large as 
the measurements suggest. It is often possible to 
refine the estimate of how much of a shape is 
present by measuring the "missing" portion. 
Portions of circles were used to record 
semicircular shapes. By multiplying the figure 
derived for the shape measured by the geometric 
fraction, actual volumes and areas can be more 
closely approximated. 

Feature Fraction 

In contrast to the descriptive nature of 
geometric fractions, feature fraction is 
interpretive. The recorder assesses the feature 
as it now exists and estimates how much is 
missing. There is a subjective element here, of 
course, but-particularly in the case of fairly 
complete features-it does not require wild 
guessing. These estimates can be used as 
confidence limits in analysis of feature 
measurements. That is, measurements for whole 
or nearly whole features can be used with 
confidence; the more of the feature that is 
missing, the more suspect the measurements 
become. In cases where it is very difficult to 
estimate how much of the feature is present, 
usually because so much of it is missing, 
assignment of a very low feature fraction flags 
the measurements as probably unreliable for 



inclusion in size assessments. Feature fraction 
values of greater than 0.7 were usually required 
for inclusion in feature dimension analysis. 

As an example of how the two fractions 
interact, consider a viga hole that is basically 
cylindrical but has a portion of the cylinder 
missing because of its location in a slanted wall. 
It might be recorded as having a geometric 
fraction (GF) of 0.8. Inspection of the edges of 
the viga hole show that the feature is still intact 
as it was used, so the feature fraction (FF) is 
1.0. The GF and the FF will be the same if it 
is apparent that the feature was originally a 
regular shape, none of which has been removed 
through deterioration. 

One cause of confusion is the tendency 
of some features (openings, especially doors) to 
become larger rather than smaller when they 
deteriorate. Whenever possible we measured 
what we considered to be the original feature. 
This was possible when, for example, two sides 
of a door remained but the other two were 
broken away; in such a case, the FF would be 
entered as 1 .0. Where we measured an opening 
that was clearly larger than the prehistoric fact, 
we could assign FF greater than 1.0. Again, as 
values approach 1.0, they are more reliable. To 
reiterate, geometric fractions describe what is 
present and help correct area and volume 
calculations; feature fractions are interpretive 
judgments of how much of the original feature is 
present and place a rating on the reliability of 
measurements. FF values close to 1.0 could be 
used to estimate full measurements of original 
features. 

Evidence for Construction 

This variable involves the straight- 
forward observation of various construction 
features. It does, however, result in some 
mixture of phenomena. That is, digging stick 
marks (Figure 3.2) are clearly evidence of part 
of the construction of a cavate. The other code 
values deal more with features that were made, 



82 CAVATE STRUCTURES 




Figure 3.2. Example of digging stick marks in the ceiling of A-10. These are especially visible 
because they cut through the smoking of the ceiling. Presumably they are evidence of 
cavate construction and/or expansion. 



such as shaping of corners or doorways. 
Although these are somewhat inconsistent, I do 
not see this as a major problem, especially in 
view of the scarcity of evidence of construction 
other than digging stick marks. Some chambers 
contain fairly clear evidence of having been 
enlarged after having been used for a while 
(Figure 3.3). Although we had no code for this 
(it was mentioned in the verbal notes), one 
might be useful in future recording. 



similar, though it can vary considerably within 
a site area, notably at Tsankawi and Group A. 
Slightly redefining this variable would have 
permitted a meaningful entry for each cavate, 
rather than lumping many cavates into an ill- 
defined "normal" group. Somewhat by default, 
"normal" in our observations means basically the 
fine-grained, white to very light gray tuff seen 
particularly in locations such as Frijoles Groups 
Fandl. 



Unusual Tuff Characteristics 

A more fruitful approach to this variable 
would have been to call it something like tuff 
type. The tuff in given areas tends to be 



Rooms Up/Rooms Down 

This observation places a room within its 
group by specifying how many rooms are up- 
canyon and how many are down-canyon to the 



RECORDING PROCEDURES 



83 




7 


■ 











Figure 3.3. Evidence for chamber expansion 
in M-10. The plaster and 
smoking of the back wall stop at 
the juncture with the right wall. 
A single coat of plaster is 
present on the right wall, while 
the back wall has several coats. 



ends of the group. It turned out to be a difficult 
blank to fill in the field, so the values were all 
entered after the field season when complete 
recording and maps were available. Counting 
rooms is far from exact in these sites because 
cliff features are subject to several 
interpretations. We made a careful count as we 
began to record each group in order to assign 
room numbers, but we invariably found that we 
had to add rooms as we studied features 
carefully during recording. Thus, counts from 
parts of groups not recorded (as at Groups A, F, 
and M) are likely to be somewhat low. In 



determining the entries for this variable, we 
treated columnar sections of cliff, so that 
multistory rooms above one another and rooms 
that opened off the back of other chambers all 
had the same number of rooms up- and down- 
canyon. This means that the counts are not a 
steady progression as one moves through the 
group unless rooms are single-level without back 
chambers. In addition to the potential for seeing 
whether location in group seems to relate to size 
and function, this observation is helpful in 
locating rooms. 

Level 

An observation that could better and 
more easily have been made in the field was 
added after we returned from the field. Each 
room was placed according to its level (or 
story). We also noted how many levels were 
present at that location. Thus each room now 
has an observation of the type "second level of 
three." In some instances intervening or 
underlying levels may have been present but not 
visible, so we assumed some other means of 
access to the upper levels. Such an assumption 
is more plausible for fully excavated cavates 
than for rooms evident only as back walls. 
Recording this attribute has the added practical 
advantage of helping to locate rooms, as does 
the rooms up/rooms down variable. This 
variable is also relevant to the possibility of 
assessing the distribution of functions and room 
sizes. 

Noncavate Data Sets 

As noted, the separate form for 
recording noncavates was a good idea that would 
have produced more useful results if it had been 
more completely compatible with the fuller 
cavate forms. Anyone using noncavate records 
individually should be aware that the first feature 
notes (rather than the base notes) often contain 
most of the observations on the noncavate; this 
is because the first feature is often most of what 
remains to be recorded. Several of these 



84 CAVATE STRUCTURES 



compatibility problems could be fixed by 
examination of the records. Here again, we 
made some later modifications to the data set. 

Feature Type 

On the field form this was a verbal 
entry, which gave rise to a variety of classes. 
The variability was reduced and the responses 
were made to follow a more consistent set of 
criteria. I would recommend that future cavate 
recording use a coded list, such as this, rather 
than verbal entries. Possible responses now 
include: 

1. Back wall. The category for which 
the noncavate form is most appropriate is the 
common instance where a masonry room was 
clearly placed against the cliff. Evidence for 
this feature includes rows of viga holes, wall 
features (niches, other holes), and plaster. In 
most cases these features involve little or no 
excavation into the cliff. 

2. Partial cavate. Features included in 
this category have at least portions of one or 
more walls in addition to the back wall. As a 
result of this definition and the nature of the 
archaeological remains, this category contains a 
wide variety of features. The use of the 
noncavate form was intended in part to cover 
remains that, for whatever reason, had little 
visible information. Thus features that might 
literally be "partial cavates" were in many cases 
recorded as cavates. Partial cavates probably 
result in about equal measure from the loss of 
fronting masonry structures from only partially 
excavated rooms and from the loss of natural 
exterior walls. Although we cannot know with 
certainty, I believe the former is probably the 
more common cause. 



There is no doubt that cavate rooms are present, 
but they are largely unmeasurable. 

4. Chamber. In a few cases the 
noncavate form was used for more or less intact 
chambers. These should probably have been 
recorded on cavate forms. They were placed on 
these forms because there was some question as 
to their prehistoric use and artificial origin. 
Because of the absence of features, however, 
loss of data and comparability was minimal. 

5. Cliff niches. These features are 
peculiar to Tsankawi (see chapter 4). They 
appear to have been outside rooms so that 
assignment of room numbers to them is 
inappropriate. They are recorded on forms both 
individually and in groups; they have all had 
their room-number values set to zero. 

6. Trails and hand-and-toe hold routes. 
There are two instances of extramural routes on 
noncavate forms from Group A. Other 
examples are present at Group I and Tsankawi, 
but they were not recorded on forms. Although 
recording these is difficult, perhaps it should 
have been done. These features also have room 
number values of zero. 

Table 3.1 shows the distribution of these 
noncavate features at the groups studied. 

Connected with Cavate 

This is filled in only when a noncavate 
and a cavate are physically linked by a feature. 
The most common example occurs where a door 
to a cavate goes through a back wall or partial 
cavate. In a few cases, a vent passes through a 
wall between a noncavate and a cavate, and 
these have been coded as connected. 



3. Filled cavate. Certain areas, 
particularly at Tsankawi and Group M, have a 
great deal of architectural and natural fill against 
the cliff base. In these areas only the top of an 
opening to a chamber is sometimes visible. 



Condition/Damage 

The first version of the noncavate form 
had only a single code for condition. During the 
recording of Group M (M-53 and after) this was 



Table 3.1. Distribution ofNoncavate Feature Types. 



RECORDING PROCEDURES 85 



Feature Type 



Group A Group F Group I Group M Tsankawi Total 



Back wall 
Partial cavate 
Filled cavate 
Chamber 

Hand & toe routes 
Cliff niches 



37 
10 


1 
2 




38 

7 











14 



2 









9 
7 
3 






27 


125 


21 


45 


2 


7 


2 


3 





2 


4 


4 



Total 



50 



45 



16 



19 



56 



186 



split into human and natural damage; we 
reconstructed the damage codes for the early 
forms from the information on the forms. 

Cavate Type 

This variable was added after the 
fieldwork to link the cavate and noncavate 
records. All cavate records received the code 
"cavate," while noncavate records were coded as 
one of the noncavate feature types listed above. 
In retrospect, it would have been useful to 
classify both "fully enclosed cavate" and 
"excavated portion of partially masonry room" 
as types of cavate, since many of the current 
noncavate "partial cavate" features are excavated 
portions of masonry rooms. The two cavate 
subtypes would have somewhat different 
information recorded, particularly for doors: 
enclosed cavates have true "exterior doors," 
while the back portions of partly masonry rooms 
would be "exterior openings." 

Collections 

Materials collected were limited to 
perishable items that seemed likely to carry 
information and to be possibly at risk of 



disintegration (see appendix 1). Two possible 
coprolites and a fused clump of corn were 
collected from Group M. A possible squash 
rind was collected from Group A. Most 
important, several bones from a very small 
infant were collected at Tsankawi. These 
remains were exposed in the disturbed fill of a 
room. Given the heavy visitation at Tsankawi, 
we decided to collect the exposed elements 
pending decisions on how best to deal with the 
vandalism of the deposits and how to most 
properly treat the burial. 

Rock Art 

Recording rock art is a time-consuming 
and specialized process. In recognition of this 
fact, we initially planned only to note the 
general size, location, and subject matter of rock 
art as features during recording. We were 
extremely fortunate in having June and Bill 
Crowder to do supplemental recording of the 
rock art. They took photographs, recorded 
locations, and compiled a summary of the rock 
art that was present in the sections where we 
worked, providing a much more detailed record 
of rock art than would have otherwise been 
available (see chapter 4). 



86 



CAVATE STRUCTURES 



Mapping 

We made a plan map for each of the 
Frijoles groups. The procedure was to establish 
Brunton compass stations, run a long tape along 
a recorded azimuth, and take right-angle 
measurements (right angles verified by Brunton) 
to edges of rooms along the line. Later Brunton 
stations were established on the line as made 
necessary by cliff curvature and topography. 
The first group we mapped was Group M, and 
we attempted to place all rooms at all levels, 
which proved time-consuming and resulted in a 
confusing map. The remaining maps measured 
only the rooms at cliff base, with higher rooms 
recorded in their relative locations, with level 
(or story) keyed. The maps locate only the 
edges of openings of rooms and do not 
reproduce floor plans. As discussed in chapter 
2, this volume includes plane table maps made 
by the Bandelier Survey crew when they are 
available. The map for Group A (Figure 2.5) 
was made by the technique described here; the 
maps for Groups F, I, and M (Figures 2.10, 
2.14, and 2.17) were made using a plane table 
and alidade by crews from the park survey. The 
plane table maps do render chamber shape, but 
not floor plan. We did not draw a plan for the 
Tsankawi group (LA 50976) because of the 
availability of Lister's map (1940b). The map 
included here is based on Lister's enlarged and 
updated map and a park survey detailed sketch 
map to show rooms added by this project 
(Figure 2.19). 

Frontal elevation or profile maps were 
drawn to compensate for the lack of measured 
plan locations (Figures 2.6, 2.11, 2.15, 2.18, 
2.20). Initially, we hoped that the NPS Branch 
of Remote Sensing would produce metrically 
correct elevations for each group; this plan was 
not followed, however. We therefore had to 
produce frontal elevations using field sketches 
and combinations of distant and close-up 
photographs. Because they attempt to render 
very irregular surfaces flat, and because of the 
perspective problems inherent in transferring 



from uncorrected photos, scales are 
approximate. Although not metrically correct, 
these elevations do show relative locations and 
shapes of rooms. The survey field crews also 
made more detailed elevation maps of Groups F, 
I, and M. The Group F survey map and 
elevation have been modified to reflect cavate 
project room numbers; the survey crews used 
cavate project numbers for the Group I and 
Group M maps. 

Individual maps were drawn for only a 
few rooms, all at Tsankawi. One of these 
rooms had a complex floor plan due to chamber 
expansion, one had many floor features revealed 
by extensive brushing and is heavily visited, and 
one contained the infant remains and some 
disturbance, as noted above. Although it would 
be desirable to have plans and profiles for each 
feature, angle and size measurements and 
complete video coverage help compensate for 
that lack. 

Photography and Video Recording 

Photographic recording fell into four 
categories. First, Bill Crowder made 
photographs (mostly 4 x 5) of the fronts of all 
the Frijoles cavates that we recorded. Detailed 
photography of cavate fronts is complicated by 
the steep slopes in front of many of the rooms, 
as well as by the high vegetation in some areas. 
Second, as noted above, the Crowders took 
photographs of each rock art panel in the areas 
recorded. Rock art photography can also be 
difficult, but through experimentation and 
darkroom techniques, most figures are visible in 
the photographs. Third, the archaeologists 
photographed specific examples of features and 
room groups. Such record photographs were 
kept to a minimum because of the complete 
videotape coverage. Lister (1940a,b) had taken 
some pictures of groups in which we worked. 
In Groups A, F, and I and at Tsankawi we took 
pictures from the same angle to show change 
after 46 years (Figures 2.7, 2.8, 2.12, 2.13, 
2.16, 2.21). Fourth, we made videotape 



RECORDING PROCEDURES 87 



recordings of each room recorded. Videotape 
has proven extremely useful in mapping of 
underwater archaeological sites, and it seemed 
likely to be useful in cavates as well. After the 
completion of standard recording in each group, 
two people (one handling camera and color, and 
one serving as coach and carrying props) 
proceeded back through each room filming each 
feature inside and out and attempting to point 
out features in each. The videotapes form a 
complete, economical, readily archived, and 
easily referenced record of the current condition 
of the cavates included in this preliminary 
project. Videotape has other advantages, as 
well. It allows detailed mapping of plaster, 
features, and rock art if such drawings become 
necessary, without the great expenditure of field 
time required to draw such maps on site. Since 
the completion of the cavate fieldwork, 
videotape technology has advanced to allow 
direct digitizing for computer mapping and 
recording of features like cavates, although this 
process is expensive. It is also possible with the 
proper equipment to produce slides from 
videotape (C. Schaafsma, personal 
communication, 1988). 

Videotape redefines the position of 
traditional still photography. If the video 
recording is of good quality, the videotape will 
hold more information than still photographs. 
But photographs still have their uses. In taking 
photographs for this project we emphasized their 
use for illustrating reports with examples of 
features and giving an idea of setting, and the 
Crowders made extensive use of them for 
recording rock art. As Bill Crowder's work 
demonstrates, photographs can be used to bring 
up detail not visible in videos and can be of far 
higher quality. Video technology is evolving 
rapidly and has advanced since 1986, but still 
photography retains its important place in 
recording, especially given the levels of 
expertise and types of equipment available to 
most archaeological projects. Photographs are 
more accessible than videotapes, since they do 
not require special equipment for viewing; 



however, they are also an archival headache. 
Since we were fairly satisfied with the results of 
videotaping, we took fewer still photographs 
than we would have done otherwise. 

Assessment of "High Tech" Techniques 

During the cavate fieldwork we used two 
tools that could, in 1986, be considered high- 
tech: direct entry of data into a laptop computer, 
and use of a video camera for recording 
features. (As noted above, the possibility of 
using remote sensing for producing elevations 
was eliminated before we even went into the 
field.) Both laptops and video cameras have 
great potential for fieldwork, and this is an 
evaluation of their usefulness as applied in 
Bandelier in 1986. 

Direct Data Entry 

Computer technology and capability are 
advancing even faster than video recording, and 
many of our experiences in 1986 could be 
avoided in 1994. Only the portions of what we 
learned that remain relevant are presented here. 
Data entry in the field has some well-known 
advantages: it circumvents a lengthy phase of 
data entry after leaving the field (time was 
projected to be short at the end of the cavate 
project field season and key-punching personnel 
scarce); it reduces keystroke errors because there 
are fewer generations of input; it creates legible 
and easily searched records; and it makes 
summary information and preliminary analysis 
available in the field. If the field workers are 
competent with the hardware and software, data 
recording can be faster than conventional 
methods. 

Our experience with field data entry was 
less than a resounding success. File transfers to 
the mainframe computer were far more difficult 
than we had been led to expect, and field 
conditions were harsher than the computer could 
easily withstand. File transfer problems result 
from the use of several unrelated software 



88 



CAVATE STRUCTURES 



packages (WordPerfect, CEO Write, Oracle) and 
staff inexperience in file translation. Use of 
compatible software and participation in field 
form creation by someone familiar with the 
program to be used in analysis will prevent these 
problems. The computer we had was a fancy, 
expensive piece of equipment--for 1986. It was 
an 8088-based machine with a single floppy 
drive, no hard drive, and 512 K of RAM. On 
the whole it worked well, although we had to be 
extremely cautious with it. The dusty and gritty 
nature of cavates would never have been so 
evident to us if we had not had the computer 
with us: the keyboard collected grit at an 
alarming rate, and in spite of careful cleaning 
after every day in the field, it was always 
noticeable. The single disk drive necessitated 
continual insertion and removal of disks, 
unavoidably introducing more dirt into the 
machine. Disks went bad at an alarming rate. 
Even today's smaller, far more capable laptop 
computers will require means of dust and grit 
protection if they are to survive and be useful in 
archaeological fieldwork. 

The actual process of data entry went 
fairly well. There was often considerable 
difficulty in reading the screen, but an adequate 
angle could usually be found. Speed of entry 
was about comparable to pencil and paper form; 
if anything, I found a tendency to take more 
complete verbal notes. We found one drawback, 
at least with the entry format we used: it was 
more difficult to check an earlier part of the 
form for previous entries and measurements than 
it was with a paper form. 

Videotape Recording 

The advantages of video recording are 
listed above. Overall, we were pleased with the 
quality of the picture. The cameras did a 
remarkable job of recording in the low light of 
many of the cavates. They also seemed to be 
less susceptible to the ubiquitous dust and grit 
than the computer, though that was again a 
concern. Of course, not every last detail of 



every wall is visible and well recorded on our 
tapes. It would probably be preferable to make 
the videotaping process part of the recording 
procedure for each room, so that all details 
could be remarked. This would considerably 
increase the recording time, however, and would 
also increase the risk of damage to the 
equipment, as it would have to be in the field 
every day instead of on selected videotaping 
days. 

As recording becomes more complex, 
there is a greater possibility that something will 
go wrong, and that is what happened here. The 
microphone of the first camera stopped working 
partway through the season, leaving large 
stretches of tape without sound. We remedied 
this by dubbing in the sound later, but the detail 
of the commentary undoubtedly suffered. This 
sort of event can considerably increase the time 
involved. On the whole, however, the video 
recording offers the advantages of completeness 
of coverage available in no other way. 
Archiving videotapes presents several problems. 
As of 1994 the NPS still did not have a 
systematic method for storing and retrieving 
tapes, although they are used by several units. 
More troubling, the technology is still new 
enough that no one knows how long images 
survive, although there is no doubt that they do 
degrade and that their storage life is probably 
quite short (less than 10 years?). To be useful 
as archival material, tapes of mid-1980s vintage 
must either be rerecorded onto longer lasting 
metal-based tape or, ideally, digitized. 
Digitization does ensure a virtually permanent, 
archival-quality record. 

Work Schedule and Field Time Spent 

The approximate distribution of time 
expenditure for this project is shown in Table 
3.2. Adjustments have been made for the 
varying hourly schedules among the people who 
worked on the project, but all days have been 
rounded down to standard work days (that is, 
some "overtime" does not show). It is quite 



RECORDING PROCEDURES 



89 



Table 3.2. Time Spent and Forms Completed by Cavate Group, 1986. 



Frijoles 



Group M 

LA 50972 



July 8-part of July 17 

field work man-hours: 236 

cavate records: 46 

noncavate records: 19 

total rooms: 140 % recorded: 46.4 

personnel: B. J. Mills and B. P. Panowski, 7Vfe days 

H. W. Toll, 6Vi days; B. Fuller, 3 (recording) 

Bill and June Crowder, 3 days each (photos & rock 

art) 



Group A 

LA 50973 



Group I 



LA 50974 



Group F 



LA 50975 



part of July 17-part of July 23, August 14-15 

fieldwork man hours: 170 

cavate records: 25 

noncavate records: 50 (48 rooms) 

total rooms: 130 % recorded: 56.2 

personnel: BJM, 4 days; BPP, 3; HWT, 5 
BF, 3 days (recording) 
JC, 3 days; BC, 2 (photos and rock art) 
E. R. Bayer, 2 days (recording) 

part of July 23-part of July 25 

fieldwork man hours: 106 

cavate records: 21 

noncavate records: 16 

total rooms: 38 % recorded: 97.4 

personnel: ERB, BJM, HWT, 2Vi days; BPP, 2 

BF, 1 day (recording) 

Crowders, 1 day each (photos and rock art) 

part of July 25-July 30 

fieldwork man hours: 126 

cavate records: 15 

noncavate records: 45 

total rooms: 106 % recorded: 56.6 

personnel: ERB, BJM, BPP, HWT, 3 days 

BF, 1 day (recording) 

Crowders, 1 day each (photos and rock art) 



90 



CAVATE STRUCTURES 



Table 3.2. (continued) 



Tsankawi 

LA 50976 



late July 30-most of August 13 

field work man hours: 268 

cavate records: 63 

noncavate records: 56 (54 rooms) 

total rooms: 117 % recorded: 100.0 

personnel: BJM, BPP, HWT, 8 days 

ERB, 1V4 days 

BF, 3 days (recording) 

Crowders, Vh days each (photos and rock art) 



clear that the groups that have more cavates than 
noncavates take considerably more time to 
record. The break-in period at Group M is also 
evident, though some of that extra time resulted 
from a more detailed mapping, and some from 
revisitation of rooms to record variables added 
after the rooms had been recorded. 

Data Manipulation 

After we returned from the field, Bruce 
Panowski coordinated a group of five people 
(including himself) who entered first the paper 
forms and then the personal computer records 
into the version of the Oracle database form that 
he had devised for the mainframe. None of 
those who entered data worked full time, and the 
process spanned a month (from August 18 to 
September 18, 1986). During this time and later 
considerable further effort was invested by Toll 
and especially Panowski in doing initial 
tabulations and locating various errors and 
omissions. After entry and checking, the files 
were transferred among several systems for 
further processing and for analysis. These 
included the University of New Mexico and NPS 
mainframe computers and personal computers. 
This section discusses various ways in which the 
data were used. Appendix 2 contains 



information on actual data set transformations 
and storage. 

Data Set Modifications and Formats 

We made some modifications to the data 
set during data correction and analysis. We 
outline these changes here for future users of the 
data and as part of the description of the data. 



Laboratory of Anthropology Numbers 

As discussed in the description of cavate 
groups, LA numbers that had been assigned to 
another project were given to the field crew. 
Replacement "official" numbers are given in 
chapter 2. In the computer data sets, the 
numbers used in the field and on all records are 
retained under the name FIELDLA, while the 
numbers on record with the Laboratory of 
Anthropology Survey Room ARMS File, are 
under the variable NEWLANO. 

Locations 

Universal Transverse Mercator (UTM) 
grid locations were assigned to all the study 
groups. It is not possible to locate every cavate 



RECORDING PROCEDURES 



91 



on 1:24,000 USGS maps, or to realistically 
specify UTM coordinates at an accuracy greater 
than 10 m. UTMs, therefore, were assigned to 
upper and lower halves of Groups A, F, and M; 
Group I was assigned a single UTM value; and 
the Tsankawi (LA 50976) group was divided 
into northern, middle, and southern groups for 
UTM assignment. Assignment of Section, 
Township, and Range location for the Tsankawi 
group was straightforward, since that entire 
location falls within a single quarter-quarter 
section. Township 18N, Range 6E, however, 
has never been platted. Moreover, the platted 
sections on the northern part of the Frijoles 
Quadrangle are somewhat irregular. Projections 
were made from the existing platted areas on the 
Frijoles Quad to the cavate locations, and 
quarter-quarter-section locations have been 
provided in the computer files for the groups 
studied. The section locations and probably the 
quarter sections are likely to be correct, but the 
locations must be recognized as projections and 
estimates only. 

Calculation of Volumes and Areas 

The measurements we took were 
designed to allow us to calculate volumes of 
chambers and areas of plane surfaces. 
Assumptions and compromises are necessary in 
calculating and using such values both because 
of the nongeometric shapes of cavates and 
because they are often only partially present. 
The procedures and formulas we used to obtain 
volume and area measurements are given below. 
An annotated version of the SAS program (SAS 
Institute 1985) is presented in appendix 2. 

Plane Shapes 

Formulas for all the plane shapes except 
the two varieties of ovals are well known and 
straightforward: 

Rectangle: Area equals width times 
length. Same formula used for bowed rectangle. 



Oval (and "natural oval"): As a 
compromise approximation, the area of oval 
features was calculated as the mean of the axes 
divided by two (giving the equivalent of a 
circle's radius), which was then squared and 
multiplied by pi. 

Trapezoid: A trapezoid may be thought 
of as a rectangle with either one or two triangles 
appended. The heights of the triangles are equal 
to the rectangle. The formula for finding the 
area is half the height times the sum of the bases 
(see, e.g., Kern and Bland 1934:vii). 

Circle: The area is the square of the 
radius times pi. Since we measured diameters 
rather than radii, the measurements were halved 
to find the radius. 

Triangle: The area is the product of the 
base and one-half the height. 

Areas were not calculated for linear or irregular 
shapes. The areas found using the above 
formulas were all multiplied by the geometric 
fraction to correct for partial and irregular 
forms. This correction is especially important in 
"round" shapes, which were often semicircular. 
We took all measurements in meters (generally 
measuring to the nearest centimeter) and 
rounded the areas only to four decimal places, 
since this allows expression of a square 
centimeter. Zeros were used for missing values 
when the data were originally entered. All areas 
of zero have been changed to missing values, as 
have all zero values for measurements, except 
for height above floor where zero is a valid 
measurement. 

Solid Shapes 

Solids are more complex to measure and 
thus involve more assumptions. Nonetheless, 
the resulting volume estimates are 
unquestionably closer to the actual volume than 
they would be if based on purely rectilinear 



92 CAVATE STRUCTURES 



measurements. The formulae used are taken 
from Kern and Bland (1934:14-34). 

Rectangular solid: The volume is the 
product of the length and width of a base and 
the height. 

Cylinder: The volume is the area of the 
base times the area of the height. This shape 
was used most often for holes in the walls, so 
that the "base" is the aperture of the hole. 
Because the area of the base requires a radius, 
the aperture diameter had to be halved. 

Hemisphere: Measuring the volume of a 
hemisphere (or a spherical segment of one base) 
requires knowing the diameter of the whole 
sphere. We assumed that the diameter taken in 
the field is the diameter of the whole sphere, 
which is unlikely to have been the case. For the 
height cf the segment the depth measurement 
was used, which is the best approximation of the 
segment height. The formula for the volume is 
one-third pi times height squared times the 
quantity three times the sphere's radius minus 
the segment's height. In contrast to our practice 
with other area and volume measurements, we 
did not use the GF on hemispheres, since the 
two measurements should approximate the 
fraction of a sphere present. This method of 
estimating hemispherical volume may give 
values somewhat greater than the actual values 
because of inflation of the radius value. Using 
this approach, we found that some volumes 
result in negative values, which shows the 
approximate nature of this calculation and 
indicates that hemispherical volumes should be 
treated with extra caution. We set volume 
values to missing for features whose 
measurements resulted in negative values (about 
12 in all). 

Truncated cone: This shape was used 
mostly for wall holes. The upper diameter was 
often difficult to measure; the cones are assumed 
to be right cones (that is, the altitude from the 
apex to the center of the base is a right angle). 



The volume formula is one-third pi times the 
height times the sum of the square of the basal 
radii and the product of the radii. 

Truncated pyramid: Several assumptions 
are active here: that the figure is the frustum of 
a right pyramid, and that the bases are 
rectilinear (and thus proportional). We 
measured two sides of the lower base of the 
figure, the height of the figure, and the length of 
the upper base. The volume formula requires 
that the area of both bases be known, which in 
turn requires that both sides of the upper base be 
known. To arrive at the width of the upper 
base, the lower base width was divided by the 
lower length, and the result was multiplied by 
the upper base length, generating the upper base 
width. The volume is then found by multiplying 
the sum of the areas of the two bases and the 
mean proportional of the two bases (the square 
root of the product of the two areas) by the 
height divided by three. 

Cone: Like the truncated cone, cones are 
assumed to be right. The volume is calculated 
as one-third pi times the basal radius squared 
times the height (basal radius found by halving 
the diameter taken in the field). 

Sphere: The volume formula is pi times 
the cube of the diameter, all divided by six. 

Pyramid: Pyramids were measured as 
four-sided right figures. The volume of such 
figures is calculated as one-third the area of the 
base times the altitude. 

In some cases we were best able to 
describe a feature as a plane figure with depth 
(usually triangular or oval). In these cases the 
area of the plane figure was multiplied by the 
depth to give a feature volume. For 
hemispheres and the "plane figures with depth," 
the GF was applied only to the area value. For 
all other shapes the volume values were also 
multiplied by the GF to correct for 
nonconformance to the geometric shape used. 



RECORDING PROCEDURES 93 



After the calculation of volumes, rectangles with 
depth were converted to rectangular solids for 
purposes of mean volume calculations. "Plane 
figures with depth" presumably do not conform 
easily to the selection of geometric solids 
because of such things as irregular depth and 
rounded corners, and this treatment of volume 
may somewhat inflate the volume derived. The 
change to solid figures was done only for 
calculation, not as a permanent change to the 
data sets. Volume values were rounded to seven 
decimal places (of a cubic meter), which allows 
expression to the nearest 10 cc. Again, field 
measurements are generally to the nearest 
centimeter, but given the difficulties and 
imprecisions in measuring volumes, the nearest 
10 cc is the closest realistic estimate, and in 
many cases even that level is probably false 
precision. 

Photographic Entries 

Four sets of photographs exist from the 
1986 cavate recording: 4x5 frontal photographs 
of individual cavates (or groups of a few closely 
spaced rooms) taken by Bill Crowder, 35 mm 
black-and-white negatives of rock art taken by 
Bill and June Crowder, color slides of rock art 
taken by the Crowders, and 35 mm black-and- 
white shots of features and cavates taken by the 
archaeologists both during the field season and 
in April 1987. Some of the photographs in each 
of these groups are of activities or are overviews 
of large portions of cavate groups. The more 
specific photographs have been entered on the 
appropriate data lines, using the numbering 
conventions listed in appendix 2. Computer 
listings of photographs sorted by feature type 
and photograph and, for base information, by 
group and cavate number are on file with the 
National Park Service. Although not every 
feature or cavate that shows in every photograph 
has been entered, the coverage should be good 
enough to permit users to locate photographs of 
examples and specific instances. Users can also 



find specific instances by referring either to 
general features, such as the wall on which a 
feature is located, or to nearby cavates as 
indexed by the frontal drawings. 

Location and Processing of Computer Data 
Sets 

The field forms for the cavate data were 
entered by various NPS personnel using the 
Oracle data base at the regional office in Santa 
Fe. After several attempts, the data were 
transported via tape to the University of New 
Mexico mainframe IBM system. Once there, 
Statistical Analysis System (SAS) formats (SAS 
Institute 1985) were generated, and the data sets 
were checked line by line and corrected as 
necessary. Much of the data analysis for this 
study was done in SAS on the university system, 
though many analyses were conducted on a 
personal computer using the Number Cruncher 
Statistical System (NCSS) and SAS-PC. During 
the project, NPS acquired SAS for its Santa Fe 
system, so we returned the data sets in SAS 
format for use in Santa Fe and did much of our 
analysis on that system. Appendix 2 lists the 
data sets and the various outputs saved and now 
in the care of NPS. 

Group Attributes 

As discussed above, we recorded two 
sets of information for each cavate and 
noncavate feature: first, general or "base" 
information, followed by recording of specific 
features. Examination of the base information 
shows the distribution of construction, condition, 
and other attributes in the various groups. 

The distribution of room types among 
groups as reflected by form types is quite 
different. Groups A and F both show strong 
predilection for rooms built against the cliff, 
while the other three groups show more frequent 
use of rooms at least partially excavated into the 
cliff (Tables 3.1, 3.3). 



94 CAVATE STRUCTURES 



Table 3.3. Room Type and Mode of Recording by Site. 



Recording Format 


Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


Cavate 


25 


15 


21 


47 


66 


174 


Noncavate 


48 


45 


16 


19 


52 


180 



Total 



73 



60 



37 



66 



118 



354 



Construction and Use 

Based on our observations, we can only 
speculate about how the cavates were actually 
made, and about whether the builders converted 
the tuff removed into building material. In 
Turkey people excavate rooms with picks, 
splitting off blocks for use in construction 
elsewhere; one man expected to spend a year 
working part-time by himself to complete a 
dwelling of five rooms and a cellar (Blair 
1970:142-145). The most common evidence of 
construction in our sample-observable in more 
than half of the cavates-is the presence of 
grooves, which are especially visible in the 
ceilings (Table 3.4, Figure 3.2). These are 
modally 2-3 cm wide, around 1 cm deep, and 
perhaps 30-40 cm long; they look very like the 
digging stick marks sometimes seen in the hard 
earth walls of pits revealed by excavation. They 
are at numerous different angles, though they 
tend to have some patterning. They look more 
like a means of final shaping of the chamber 
than a means of removal of large chunks. 

At least partly because of different tuff 
types, digging stick marks are less common at 
Tsankawi, but at both Tsankawi and Frijoles it 
was common practice to smooth the walls of a 
cavate as high as the plaster was to go (often 
around 1 m) and leave the upper wall and the 
ceiling rough. Remodeling and housekeeping 
seem not to have included removing the smoked 
layer of tuff from the ceiling; replastering the 
walls seems to have sufficed. In contrast, recent 



Turkish occupants of cavate rooms cut into tuff 
similar in appearance to that in Frijoles, remove 
smoke-blackened rock from the ceilings with 
hammers, and replaster on an annual basis 
(Riboud 1958:138-139). Some form of grinding 
or polishing similar to that used to smooth lower 
walls probably also finished the shaped openings 
and doors, but we noted no evidence other than 
smooth, regular surfaces. As far as we could 
determine, 95 percent of the cavates we 
examined resulted more from artificial 
excavation of the cliff than from use of natural 
pockets (Table 3.5). The use of masonry was 
the least at Tsankawi (Table 3.6). Among the 
Frijoles groups, there was less use of masonry in 
upper Group M than in the other groups, though 
M-60 contains one of the most substantial 
remaining walls recorded. Masonry remains in 
about 20 percent of the cavate features of 
Groups A, F, and I; if the upper part of Group 
F had also been recorded, Group F would have 
been likely to stand out as containing more 
masonry than the other groups, because of the 
number of masonry features and dividing walls 
present there. 

The greatest use of multilevel rooms 
occurs at Group M, where a large central cluster 
of cavate rooms is located above a substantial 
house mound. Group F has a similar high 
central cluster, but it forms a smaller percentage 
of the rooms recorded. There are also many 
multilevel rooms at Group A, though upper 
Group A contained no areas with four levels. The 
part of A we did not record contains further 



Table 3.4. Evidence for Construction. 



RECORDING PROCEDURES 95 



Type of Evidence 



Group A Group F Group I Group M Tsankawi Total 



Cavates 
Evidence lacking 
2: Digging stick marks 
3: Shaped opening 
4:2 + 3 

5: Shaped corners 
6:2 + 5 
7: 2 + 3 + 5 
8: Floor leveling 



4 
3 

2 
4 
7 
4 
1 



1 


3 


6 


3 








1 


1 


1 


3 


1 


6 


5 


5 









15 


24 


47 


4 


13 


29 


2 


5 


7 


6 


8 


18 


13 


3 


24 


6 


6 


26 


1 


7 


22 








1 



Total 



25 



15 



21 



47 



66 



174 



Total digging stick marks 



16 



13 



15 



17 



30 



91 



Evidence lacking 


43 


36 


14 


18 


51 


162 


Digging stick marks 


2 


9 


1 





4 


16 


Shaped opening 











1 





1 


Shaped corners 


4 





1 








6 



Total 



49 



45 



16 



19 



55 



185 



Combined total 



74 



60 



37 



66 



121 



359 



96 



CAVATE STRUCTURES 



Table 3.5. Estimate of Excavated versus Natural Space (Cavates Only). 



Portion Excavated 



Group A Group F Group I Group M Tsankawi Total 



Completely excavated 


12 


9 


13 


20 


41 


95 


> Half excavated 


10 


6 


7 


26 


22 


71 


> Half natural 


2 





1 





2 


5 


Completely natural 


1 








1 


1 


3 



Total 



25 



15 



21 



47 



66 



174 



multilevel areas, and it is likely that of the five 
groups, Group A has the highest frequency of 
stacked rooms. Although the percentage of 
upper-level rooms is somewhat less at Group I, 
it still far outstrips that at Tsankawi (Table 3.7). 
The stepped nature of the tuff outcrops at 
Tsankawi made it more difficult for us to assign 
levels, and in most areas it precluded 
construction of more than one or two levels. 
Tsankawi has a markedly lower frequency of 
rooms built at least partly above other rooms. 



Fill 



The paucity of fill in a great majority of 
the cavates is evident in Tables 3.8 and 3.9: 
one-fifth were found to have no fill at all. Most 
of those containing fill in the Frijoles groups 
contain a mixture of disintegrated tuff and a very 
fine, floury dust, presumably mostly aeolian in 
origin. There is frequently some mixture of 
organic materials, such as grass and leaves, 
which often look as though they may also have 
blown in or may have been brought in by 
rodents. Animal dung was sometimes present in 
Frijoles, but never in quantity. In the lowermost 
tier of rooms at Tsankawi, several rooms had a 
thick, hard layer of dung. This tier of structures 
also had many of the most deeply filled rooms 
because of its location at the base of a low cliff. 
The few deeply filled rooms in Frijoles are next 
to the rockfall in Group I or have openings 



below the level of the exterior ground surface so 
that fill runs into the structure. A few chambers 
of which only the top few centimeters were 
visible were recorded on noncavate forms at 
Tsankawi and Group M, so that the deeply filled 
chamber count in Table 3.9 is somewhat 
depressed. The consistency with which cavate 
rooms with raised entries have very little fill 
leads me to suspect that they may never have 
had much. Alternatively, earlier visitors may 
have been very thorough in their "investigations" 
of these features. 

Small fill amounts make it difficult to 
detect whether there has been disturbance of the 
fill, because every entry into a room with a few 
centimeters of fill is a substantial disturbance. 
Our forms did not have a specific entry for 
noting the presence of pothunting activity, and 
generally we saw little evidence of it. Two of 
the deeply filled, red tuff rooms at Tsankawi do 
have major "potholes" in them (one exposing the 
bones of an infant). One or two Group F rooms 
have dirt and roots clinging to the walls, 
suggesting that fill has been removed, and 
another has had its floor cut through with metal 
tools. 

The few rooms we examined that were 
especially difficult to get into did have more 
cultural material in them. Four smaller, higher 
rooms had quantities of chunks of tuff in them 



RECORDING PROCEDURES 97 



Table 3.6. Masonry Presence and Type. 



Masonry Type 


Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


Cavates 














Masonry absent 


20 


12 


17 


43 


63 


155 


Large blocks, simple 


4 





4 


4 


1 


13 


Small rock, much mortar 


1 


2 











3 


Coursing absent 





1 








2 


3 


Total 


25 


15 


21 


47 


66 


174 


Percent with masonry 


20.0 


20.0 


19.0 


8.5 


4.5 


10.9 


Non-cavates 














Masonry absent 


47 


44 


15 


19 


56 


181 


Large blocks, simple 


1 





1 








2 


Small rock, much mortar 





1 











1 


Coursing absent 


1 














1 



Total 



49 



45 



16 



19 



56 



185 



Combined total 

Percent of group 
with masonry 


74 60 
9.5 6.7 


37 66 
13.5 6.1 


122 
2.5 


359 
6.4 


along with relatively hi 


gh artifact frequencies 


Natural Features 







(rooms A-4, A-57, F-56, and M-15). The 
likeliest explanations for the unusual rock fill are 
either the collapse of portions of walls or 
"basketball" practice after abandonment, 
probably during the last hundred years. 



The greater variability in tuff found at 
Group A and Tsankawi is shown in Table 3.10. 
These assessments of tuff are somewhat 
impressionistic. Still, the tuff at Group M 



98 CAVATE STRUCTURES 



Table 3.7. Room Level, Cavates and Noncavates Combined. 



Level 



Group A Group F Group I Group M Tsankawi 



Total 



First 
Second 
Third 
Fourth 
Total 



38 


34 


27 


29 


107 


235 


29 


20 


9 


22 


11 


91 


6 


4 


1 


12 





23 





2 





2 





4 


73 


60 


37 


65 


118 


353 



Table 3.8. Fill Type, Cavates Only. 



Fill Type 



Group A Group F Group I Group M Tsankawi 



Table 3.9. Fill Depth, Cavates Only. 



Total 



Clear floors 


7 


2 


2 


16 


8 


35 


Disintegrated tuff 


7 


3 


4 


18 


5 


37 


Dung and tuff 





1 








1 


2 


Aeolian/alluvial 


7 


3 


10 


7 


20 


47 


High organic content 








1 


1 





2 


Tuff, dung, organic 





3 


3 


3 


11 


20 


Aeolian/alluvial, 


4 


2 








21 


27 


tuff, organic 














Total 


25 


14 


20 


45 


66 


170 



Fill Depth 


Group A 


Group F 


Group I 


Group 


M 


Tsankawi 


Total 


< 10 cm 


22 


6 


11 


38 




28 


105 


10-25 cm 


1 


4 


6 


4 




17 


32 


25.1-50 cm 








2 







12 


14 


> 50 cm 





1 


1 







6 


8 


Total 


23 


11 


20 


42 




63 


159 



RECORDING PROCEDURES 99 



Table 3.10. Tuff Characteristics, Cavates Only. 



Tuff Type 


Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


Ordinary white 


12 


15 


19 


23 


18 


87 


Softer than usual 











1 





1 


Stratified, variable 


9 





1 





1 


11 


Large fibrous chunks 


4 





1 


23 


1 


29 


Highly porous 














24 


24 


Soft red 














22 


22 



Total 



25 



15 



21 



47 



66 



174 



contains abundant soft, fibrous chunks of tephra, 
which probably leads to an increased rate of 
deterioration. 

The most common nonhuman users of 
the cavates are insects (Table 3.11). Wasp nests 
are probably the most common evidence of 
insects, and larval casings are also found. A bee 
swarm was noted at Group A, but it was happily 
in a natural hole well above where we were 
working. A single room in Group M was 
almost completely plastered with some form of 
larval casing that we did not see in any other 
place. We saw no rodent nests, but tracks and 
scat were commonplace. Pack rats seemed to be 
frequenting only two features at Tsankawi and 
one at Group A. The absence of bats was 
surprising, but the shallowness of many cavates, 
and the rounded, perchless nature of most of the 
ceilings may account for the bats seeking the 
higher cracks rather than the cavates. Birds 
seem to be primarily casual, though not 
infrequent, visitors to the cavates, except for the 
vultures that use the low, dark, dank, mostly 
filled rooms at one end of Group I. We watched 
with bemused indignation as a rock wren 
brazenly removed our room tags from Group M 
very shortly after we had put them in place. 
Cattle and sheep have made the greatest impact 
on cavates, but fortunately their use is fairly 



limited. Livestock (and perhaps wild burros or 
even deer) has entered many cavates, especially 
at Tsankawi. Where livestock has been 
abundant, a thick layer of dung is present, which 
may well protect room fill; heavy use by large 
animals has apparently worn grooves 20-30 cm 
wide ("incised dados") into cavate walls just 
above the top of the fill. We recorded four 
instances of these grooves at Tsankawi and one 
at Group M. "Combination" use, as noted on 
our forms, generally meant use by several of the 
common categories of nonhuman users, 
especially insects, rodents, and ungulates. We 
noticed no visitation or use by carnivores, 
though we found some tracks in Group I that we 
could not identify. Perhaps incorrectly, we did 
not include pothunters as nonhuman users. 

Cavate Condition 

On the whole, extremely fragile cavates 
are likely to have fallen apart long ago, 
becoming, for our purposes, noncavates or 
nonexistent. In comparing the estimated stability 
of cavates and noncavates, we see that relatively 
more cavates seemed to be stable and more that 
seemed to be facing major problems, while a 
higher percentage of the noncavates fall in the 
middle ranges (Table 3.12). Noncavates tend to 
be more exposed and either have fewer features 



100 



CAVATE STRUCTURES 



Table 3.11. Nonhuman Use ofCavates. 



Animal Use 


Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


None visible 


10 


3 


4 


21 


30 


68 


Ungulates 





2 


2 





10 


14 


Pack rats 


1 











2 


3 


Insects 


7 


4 


6 


15 


9 


41 


Combinations 


5 


3 


3 


6 


8 


25 


Raptors and vultures 








4 








4 


Other birds 


1 


1 





5 


1 


8 


Other rodents 


1 


2 


2 





6 


11 


Total 


25 


15 


21 


47 


66 


174 



Table 3.12. Overall Stability of Ca votes and Noncavates. 



Degree of Threat 


Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


Cavates 














Apparently stable 


4 


1 


5 


8 


13 


31 


Lesser threat 


8 


9 


6 


9 


35 


67 


Greater threat 


7 


4 


6 


17 


13 


47 


Major problem 


5 


1 


3 


7 


4 


20 


Total 


24 


15 


20 


41 


65 


165 


Noncavates 














Apparently stable 


3 





1 


1 


7 


12 


Lesser threat 


13 


22 


7 


1 


41 


84 


Greater threat 


25 


22 


8 


10 


7 


72 


Major problem 


7 


1 





7 


1 


16 


Total 


48 


45 


16 


19 


56 


184 


Combined total 


72 


60 


36 


60 


121 


349 



RECORDING PROCEDURES 



101 



at risk (as with back walls) or have already 
experienced most of what loss was possible. 
The majority of both types falls into the middle 
ranges. Appendix 3 lists all cavates and 
noncavates by their apparent stability and 
includes the notes for each room considered to 
have an imminent possibility of collapse. In 
spite of its greater visitation, Tsankawi seemed 
to have the highest percentage of stable and 
"lesser threat" cavates, while Group M seemed 
to have the greatest deterioration problem. Of 
course, Group M has a large number of partial 
rooms, which accounts in part for its lower 
stability (see Table 3.12). Tsankawi and Group 
F have the highest relative frequencies of 
complete or mostly complete cavate chambers, 
with the absolute number being by far the 
greatest at Tsankawi. 

The amount of graffiti in noncavates is 
really quite small (Table 3.13), but complete 
cavates attract far more attention than do back 
walls. Of the areas in which we worked, Group 
A is the most seriously affected by graffiti, 
which is reflected in both the cavate and the 
noncavate data. Signs of serious wear are most 
evident at Tsankawi. The exposed nature of 
noncavates is reiterated in the assessment of 
natural impacts presented in Table 3.14, where 
severe weathering combined with severe cliff 
deterioration is the most common category. 

The presence of natural and human 
damage was assessed on two levels. First, we 
estimated the overall stability of each cavate and 
noncavate; for each cavate structural feature 
(wall, floor, ceiling), we recorded human and 
natural damage, and we used the same codes for 
an overall assessment of each noncavate. Tables 
3.13 and 3.14 summarize this information by 
group and by chamber location. 

Of the groups we studied in 1986, 
Tsankawi has the most complex situation 
regarding the condition of the rooms. The 
chambers at Tsankawi are visited daily, 
sometimes by rather large numbers of people, 



while access to Groups A, F, I, and M is limited 
to varying degrees. In the two weeks when we 
worked at LA 50976 on a regular basis, one 
substantial portion of shaped tuff was broken off 
in the area of TS-57 and TS-60, campers were 
present in TS-57, and numerous rocks were 
thrown and dislodged in the midst of the room 
concentration (some too close for our comfort). 
In rooms where we swept away shallow fill, a 
very clear wear pattern was present: the plaster 
floors are missing completely around the door 
and into the middle of the room, while 
aboriginal floor, sometimes in remarkably good 
condition, is present around the edges. Even 
more remarkable is the presence of intact loom 
anchors in some of the rooms in spite of their 
being covered by only a few centimeters of fine 
fill. The loom anchors seem to be on the 
margin of the worn area, probably because the 
looms were placed at the edge of the prehistoric 
traffic flow, as well. 

On days of high traffic and bad 
behavior, complete closure of the site seemed 
the best course. On the other hand, many of the 
visitors to this area genuinely appreciated the 
opportunity to see these structures up close. 
Powers favors importing fill for shallowly filled 
rooms, but while this would help protect 
remaining floor features, it would be difficult to 
do and would have its own impacts. Finding fill 
that is not part of an archaeological deposit will 
require going to the canyon bottom, and 
transporting it will be laborious and potentially 
damaging to the structures. Closure of 
especially fragile areas, such as the TS-54-TS-66 
area, has some appeal, but closure would only 
keep out law-abiding people. The best 
alternative may be to carefully, systematically, 
and completely clear ("excavate," if you will) 
shallowly filled rooms, collect samples, fully 
record the floors, and then refill them to the 
extent possible. This should not be a hurried, 
under recorded, and under reported salvage job. 
Rather, it should receive the full excavation 
status warranted by its location in a national 
monument. Such a project would not be cheap, 



102 CAVATE STRUCTURES 



Table 3.13. Human Damage by Group and Chamber Location. 



A. Human Damage to Cavate Walls, Floors, Ceilings 



Type of Damage 


Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


No damage 


74 


48 


74 


144 


197 


537 


Minor graffiti 


13 


15 


12 


17 


19 


76 


Major graffiti 


19 


2 


6 





4 


31 


Graffiti & other 


1 








1 


1 


3 


Obvious wear 


6 


4 


2 


2 


17 


31 


Tourist blasting 














2 


2 


Minor vandalism 


4 


1 





3 


3 


11 


Major vandalism 














3 


3 


Total 


117 


70 


94 


167 


246 


694 



B. Human Damage to Noncavates 



Type of Damage 


Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


No damage 3 


42 


12 


16 


11 


2 


83 


Minor graffiti 


3 


1 





1 





5 


Major graffiti 


1 














1 


Obvious wear 














1 


1 


Tourist blasting 














3 


3 


Minor vandalism 


1 














1 


Major vandalism 


1 














1 


Total 


48 


13 


16 


12 


6 


95 



RECORDING PROCEDURES 



103 



Table 3.13. (continued) 



C. Human Damage by Chamber Location, Cavates Only h 





Right 


Back 


Left 


Exterior 








Type of Damage 


Wall 


Wall 


Wall 


Wall 


Floor 


Ceiling 


Total 


No damage 


112 


96 


109 


60 


50 


109 


537 


Minor graffiti 


23 


18 


22 


8 





5 


76 


Major graffiti 


9 


13 


5 


3 





1 


31 


Graffiti & other 





1 


1 


1 








3 


Obvious wear 


2 


2 


5 


11 


10 


1 


31 


Tourist blasting 





2 














2 


Minor vandalism 


1 


7 


2 








1 


11 


Major vandalism 


1 


1 


1 











3 



Total 



148 



140 



145 



83 



60 



117 



694 



a In 91 cases there was no entry; probably no damage. 

b Cavate structural features only; 14 features with damage not recorded are not shown. 



but it would likely be no more expensive than 
some of the other alternatives. It is also a 
sensible plan in archaeological terms, given the 
current visitation regime. 

Another impact that is more severe at 
Tsankawi than at the other sites is the effect of 
livestock. The lower, red tuff rooms were 
heavily used by sheep and/or cattle; combined 
with the friable nature of this geological unit, 
the livestock use has ensured that few wall 
features remain in the majority of the red tuff 
rooms. On the whole these low rooms have 
mon fill than most cavates, which may have 
preserved features closer to the floors. 



The rate at which cavates deteriorate can 
only be discussed generally, given the lack of 
any precise baselines. Lister's photographs from 
the late 1930s allow us to make some 
comparisons (Figures 2.7, 2.8, 2.12, 2.13, 2.16, 
2.21). They show clearly that the elements most 
at risk are masonry: all the major changes 
visible involved masonry collapse, in spite of the 
attempts to slow that process through 
stabilization. On the whole, it seems that there 
is very slow (on a human scale) erosion of 
natural features, punctuated by occasional, much 
larger but relatively infrequent events, such as 
the rockfall at Group I, the dislodged blocks at 
Group A and Tsankawi (A-72 and TS-22), and 



104 CAVATE STRUCTURES 



Table 3.14. Natural Damage by Group and Chamber Location. 



Type of Damage 



A. Natural Damage to Cavate Walls, Floors, Ceilings 
Group A Group F Group I Group M Tsankawi 



Total 



No damage 


19 


10 


17 


28 


37 


HI 


Moderate erosion 


21 


16 


7 


31 


35 


110 


Severe erosion 


4 


2 


3 


10 


9 


28 


Moderate cliff 


15 


4 


2 


16 


18 


55 


deterioration 














Severe cliff 


7 


5 


3 


18 


9 


42 


deterioration 














Moderate erosion & 


8 


16 


17 


12 


66 


119 


deterioration 














Severe erosion & 


43 


17 


45 


52 


74 


231 


deterioration 














Total 


117 


70 


94 


167 


248 


696 



B. Natural Damage to Noncavates, Including Nonrooms 



Type of Damage 


Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


Absent 


1 








1 





2 


Moderate weathering 


1 


1 








4 


6 


Severe weathering 


2 


5 


2 


6 


8 


23 


Severe deterioration 


2 








1 





3 


Moderate weathering 
& deterioration 


5 


1 


1 





13 


20 


Severe weathering & 
deterioration 


37 


38 


13 


10 


31 


129 


Total 


48 


45 


16 


18 


56 


183 



RECORDING PROCEDURES 



105 



Table 3.14. (continued) 



C . Natural Damage by Chamber Location, Cavates Only 





Right 


Back 


Left 


Exterior 




Type of Damage 


Wall 


Wall 


Wall 


Wall 


Flo( 


No damage 


25 


34 


22 


8 


3 


Moderate erosion 


25 


29 


27 


10 


4 


Severe erosion 


5 


6 


4 


4 


5 


Moderate cliff 


15 


12 


12 


4 


2 


deterioration 












Severe cliff 


7 


6 


7 


6 


7 


deterioration 












Moderate erosion 


26 


27 


22 


17 


7 


& deterioration 












Severe erosion & 


45 


27 


51 


34 


32 


deterioration 













Ceiling Total 



19 


111 


15 


110 


4 


28 


10 


55 


9 


42 


19 


118 


42 


231 



Total 



148 



141 



145 



83 



60 



118 



695 



fallen boulders at Tsankawi with evidence of 
former rooms. Robert Preucel, formerly of the 
PARP team, reports that between 1980 and 1985 
three cavates collapsed in the Garcia Canyon 
area, presumably due to natural processes such 
as freezing and thawing (R. Preucel, personal 
communication, 1988). 



Human visitation compounds natural 
erosion and, worse, causes disturbance of 
deposits and features that are subject to little 
natural damage. The human threat is more 
focused and destructive to archaeological 
information than other types of damage. At 
least in theory, however, it is also more easily averted. 



Cavate and Noncavate Features: 
Definitions, Distributions, and Dimensions 



This chapter defines each feature type 
recorded in 1986, as well as presenting 
occurrence and metric data for each. This 
information forms the basis for the analyses, 
which use selected features and attributes to look 
for patterning within and among the various 
cavate groups. All measurements are in meters, 
though in the discussions we sometimes refer to 
centimeters or other smaller units. Volume 
measurements are given to five decimal places, 
which might appear to be a pretension to 
extraordinary precision. The field measurements 
were taken to the nearest centimeter, however, 
and 1 cc is 0.000001 m 3 , so that technically the 
data as presented drop one decimal point of 
precision. In interpreting all the measurements, 
and especially the volumes, one must remember 
that measuring cavate features usually requires 
fitting complex, curvilinear shapes into 
simplified geometric shapes. Thus, most of the 
measurements are only best approximations. In 
some cases, measurements only from features 
thought to be reasonably (more than 70-75 
percent) complete, or feature heights only in 
chambers where fill is minimal (15-20 cm is the 
usual cutoff) are used. In other cases, 
combining similar shapes gives larger samples 
and better overall characterization of feature 
size. Wherever these screens and combinations 
have been applied, the tables are annotated 
accordingly. 



In examining distributions of features 
across cavate groups and locations within 
chambers, the varying numbers of observable 
cases from which the sample is drawn are 
obviously crucial to understanding what is 
normal and what is unusual. As discussed in 
chapter 2, different groups have different 
compositions, and this can be seen in Table 4.1. 
Although the gross number of features recorded 
can generally be ranked in descending order as 
Tsankawi, Group A, Group M, Group F, and 
Group I, the pattern certainly does not hold for 
all categories. The standard of comparison 
cannot always be number of cavates, partly 
because of the variability in cavates. Further, 
while some features can occur only in fairly 
complete cavates, other features are found in 
both cavates and noncavates, and still others, 
such as viga holes, are more likely to occur in 
noncavates than cavates. As discussed in 
chapter 3, the noncavate form allowed us to 
record groups of features and size ranges for 
those groups. The metric data presented here 
are only for individually recorded features. For 
each feature type containing examples recorded 
as groups, the distribution table includes an 
entry for "noncavate, grouped" showing the 
numbers of features so recorded at each cavate 
group. A total of 362 features were group- 
recorded, and 83 percent of these fall into three 
feature types: 234 viga holes (65 percent), 37 



107 



108 CAVATE STRUCTURES 



Table 4.1. Overall Occurrence ofCavates, Noncavates, and Features by Group. 



Group A Group F Group I Group M Tsankawi 



Total 



Total rooms recorded 


73 


60 


37 


65 


118 


353 


Row percentage 


20.7 


17.0 


10.5 


18.4 


33.4 




Cavates recorded 


25 


15 


21 


46 


65 


172 


Row percentage 


14.5 


8.7 


12.2 


26.7 


37.8 




Noncavates recorded 


48 


45 


16 


19 


53 


181 


Row percentage 


26.5 


24.9 


8.8 


10.5 


29.3 




Total features 


667 


402 


318 


616 


1394 


3397 


Row percentage 


19.6 


11.8 


9.4 


18.1 


41.0 




Total walls recorded 


137 


104 


89 


123 


265 


718 


Row percentage 


19.1 


14.5 


12.4 


17.1 


36.9 





indeterminate holes (10 percent), and 29 hand- 
and-toe holds (8 percent). 

The 53 feature types recorded are 
discussed individually in the following pages 
(Table 4.2). They are grouped into the 
following categories: structural features, such as 
chambers, doors, and walls; floor features; wall 
features; and rock art. Except for feature types 
of which only a few examples were found, the 
occurrence across groups, the shape 
distributions, and appropriate metric data are 
tabulated for each type. Particularly where 
classifications were difficult or subjective-as in 
the case of holes in walls, for example- 
comparisons across types have also been made. 
This chapter includes some discussion of co- 
occurrence of functionally related feature types, 
and chapter 5 further explores feature 
associations. 

Structural Features 

Chamber (Code 1) 

This feature type was used to give the 
best approximation of the entire space enclosed 



by a cavate. The components enclosing the 
space (walls, floor, ceiling) were also recorded, 
but this feature type is best suited to describing 
the room as a whole. 

The shape-by-group frequencies show 
whole chambers that were recorded either as 
single shapes (most cases) or as combined 
shapes (Table 4.3). The combined shapes are 
shown as the base shape plus the top shape; one 
chamber at Tsankawi resulted from removing the 
wall between two truncated-pyramid chambers. 
The single cylindrical "chamber" is a partial 
cavate at Tsankawi, of which only the base was 
recordable. A clear trend toward rounded 
shapes is seen at Tsankawi, though the most 
common shape overall, the truncated pyramid, is 
also abundant there. Of the 182 cases shown, 
166 were recorded as cavate features and 17 as 
noncavates (1 1 partial cavates, 4 filled cavates, 
and 2 chambers). 

The mean volume of chambers is also 
larger at Tsankawi, but variability in chamber 
size is also much greater there. Thirteen 
chambers at Tsankawi have volumes greater than 
8 m 3 , which is larger than the largest for any 



FEATURES 



109 



Table 4.2. Occurrence of All Individually Recorded Features in Cavates and Noncavates, by 
Cavate Group (Number and Percentage). 



Feature 


Group A 
(n) 
(%) 


Group F 
(n) 
(%) 


Group I 
(n) 
(%) 


Group M 
(n) 
(%) 


Tsankawi 
(n) 
(%) 


Total 
(n) 

(%) 


Structural features 














Chamber 


26 
13.5 


18 
9.3 


22 
11.4 


48 
24.9 


79 
40.9 


193 
5.7 


Exterior door 


10 
15.4 


11 
16.9 


6 
9.2 


6 
9.2 


32 
49.2 


65 
1.9 


Exterior opening 


16 
22.2 


5 
6.9 


2 
2.8 


10 
13.9 


39 

54.2 


72 
2.1 


Interior door 


4 
10.3 





8 
20.5 


8 
20.5 


19 
48.7 


39 
1.1 


Passage 


1 
33.3 








1 

33.3 


1 
33.3 


3 
0.1 


Natural wall 


137 
19.0 


104 

14.4 


89 
12.4 


124 
17.2 


266 
36.9 


720 
21.2 


Masonry wall 


1 
16.7 





3 
50.0 


2 
33.3 





6 
0.2 


Ceiling 


20 
16.4 


12 
9.8 


14 
11.5 


28 
23.0 


48 
39.3 


122 
3.6 


Masonry and tuff wall 


3 
100.0 














3 
0.1 


Chamber corner 


33 
22.0 


18 
12.0 


22 
14.7 


14 
9.3 


63 
42.0 


150 
4.4 


Compass location point 














14 
100.0 


14 
0.4 


Floor features 














Floor 


18 

27.7 


3 
4.6 


5 

7.7 


31 
47.7 


8 
12.3 


65 
1.9 


Firepit 


12 
42.9 


1 
3.6 


3 
10.7 


7 
25.0 


5 
17.9 


28 
0.8 


Floor burn 


1 
16.7 








4 
66.7 


1 
16.7 


6 

0.2 


Subfloor pit 


4 
19.0 





2 
9.5 


6 
28.6 


9 
42.9 


21 
0.6 


Floor depression 


18 

27.7 


3 
4.6 


5 
7.7 


31 
47.7 


8 
12.3 


65 
1.9 



110 CAV ATE STRUCTURES 



Table 4.2. (continued) 



Feature 


Group A 
(n) 
(%) 


Group F 
(n) 
(%) 


Group I 
(n) 
(%) 


Group M 
(n) 
(%) 


Tsankawi 

(n) 
(%) 


Total 
(n) 

(%) 


Floor pit complex 





1 
25.0 








3 
75.0 


4 
0.1 


Posthole 














9 
100.0 


9 
0.3 


Floor ridge 


3 

27.3 








7 
63.6 


1 
9.1 


11 
0.3 


Metate rest 





1 


1 


4 





6 






16.7 


16.7 


66.7 




0.2 


Loom anchor 








1 


12 


16 


29 








3.4 


41.4 


55.2 


0.9 


Step 








1 

33.3 


2 
66.7 





3 
0.1 


Axe groove 














5 
100.0 


5 
0.1 


Adobe collar 





1 
100.0 











1 
0.0 


Deflector 








1 
33.3 





2 
66.7 


3 
0.3 


Wall features 














Large floor-level niche 


23 


25 


8 


24 


48 


128 




18.0 


19.5 


6.3 


18.8 


37.5 


3.8 


Wall niche 


29 


28 


9 


26 


64 


156 




18.6 


17.9 


5.8 


16.7 


41.0 


4.6 


Slot 


3 


1 


1 


5 





10 




30.0 


10.0 


10.0 


50.0 




0.3 


Viga hole 


97 

34.4 


37 
13.1 


19 
6.7 


55 
19.5 


74 
26.2 


282 
8.3 


Possible latilla hole 


21 


5 





6 


16 


48 




43.8 


10.4 




12.5 


33.3 


1.4 


Beam seat 


39 


7 


9 


45 


76 


176 




22.2 


4.0 


5.1 


25.6 


43.2 


5.2 


Indeterminate hole 


91 


78 


51 


74 


277 


571 




15.9 


13.7 


8.9 


13.0 


48.5 


16.8 


Possible loom support 


1 
3.6 





3 
10.7 


3 
10.7 


21 
75.0 


28 
0.8 



Table 4.2. (continued) 



FEATURES 111 



Feature 


Group A 
(n) 
(%) 


Group F 
(n) 
(%) 


Group I 
(n) 

(%) 


Group M 
(n) 

(%) 


Tsankawi 
(n) 
(%) 


Total 
(n) 
{%) 


Smokehole 


12 
20.7 


11 
19.0 


8 
13.8 


2 
3.4 


25 
43.1 


58 

1.7 


Vent 


5 
15.6 


8 
25.0 


4 
12.5 


3 
9.4 


12 
37.5 


32 
0.9 


Groove 


2 
25.0 





1 
12.5 


1 
12.5 


4 
50.0 


8 
0.2 


Wall depression 


24 
35.3 


12 
17.6 


6 
8.8 


7 
10.3 


19 

27.9 


68 
2.0 


Wall ledge 


4 
36.4 


1 
9.1 


2 
18.2 


1 
9.1 


3 
27.3 


11 
0.3 


Vertical ceiling hole 


1 
1.8 











55 
98.2 


56 
1.6 


Narrow wall incisions 


1 
4.0 











24 
96.0 


25 
0.7 


Hand-and-toe hold 











4 
100.0 





4 
0.1 


Incised dado 











1 

20.0 


4 
80.0 


5 
0.1 


Cliff niche 














5 
100.0 


5 
0.1 


Rock art 














Geometric petroglyph 


2 
12.5 


4 
25.0 





4 
25.0 


6 

37.5 


16 
0.5 


Geometric pictograph 


2 
28.6 





4 
57.1 


1 

14.3 





7 
0.2 


Zoomorphic petroglyph 


1 
3.1 


4 
12.5 


6 
18.8 


9 
28.1 


12 
37.5 


32 
0.9 


Zoomorphic pictograph 











2 
40.0 


3 
60.0 


5 
0.1 


Indeterminate 
petroglyph 


3 
9.7 


4 
12.9 


4 
12.9 


8 
25.8 


12 
38.7 


31 
0.9 


Indeterminate 
pictograph 


8 
36.4 


1 
4.5 


1 
4.5 


7 
31.8 


5 
22.7 


22 
0.6 


Handprint 











3 
100.0 





3 
0.1 



112 CAVATE STRUCTURES 



Table 4.2. (continued) 



Feature 


Group A 
(n) 

(%) 


Group F 
(n) 
(%) 


Group I 
(n) 

(%) 


Group M 
(n) 
(%) 


Tsankawi 
(n) 
(%) 


Total 
(n) 
(%) 


Anthropomorphic 
petroglyph 





1 
7.1 


2 
14.3 


4 
28.6 


7 
50.0 


14 
0.4 


Total 


665 


401 


318 


616 


1395 


3395 


Percent 


19.6 


11.8 


9.4 


18.1 


41.1 


100.0 



Note: Row percentages are shown for each feature, and percentages of the total count are shown in 
the "Total" column and row. 



Frijoles group (see Figures 4. 1-4. 3b). At least 
two very large chambers at Tsankawi (TS-36 
and TS-66) resulted from the prehistoric removal 
of walls dividing what once were two separate 
chambers. Six of the rooms greater than 8 m 3 , 
and three of the five largest (larger than 13 m 3 ), 
are located in the top tuff stratum, which 
suggests that that location was favorable to large 
chamber construction (TS-25, TS-26, TS-15, 
TS-27, TS-24, TS-20 in ascending order of size; 
see Figures 2.19, 2.20). The vesicular nature of 
the tuff there may have provided natural 
beginnings for rooms. In addition, the tuff in 
the top stratum seems to be indurated, perhaps 
making the chambers more stable. Three of the 
large chambers are located in the soft, red, 
lowermost tuff layer (TS^4, TS-29, and TS-36), 
and all are less than 10 m 3 . TS-36, an expanded 
chamber, contains much fill and its volume may 
have been underestimated. The remaining four 
large chambers are in and adjacent to the 
concentration of structures in the middle levels 
of the group (TS-50, TS-66, TS-64, and TS-59). 
Both TS-50 and TS-66 have two doors, and TS- 
59, TS-64, and TS-66 contain diverse features 
and rock art. The largest chamber (TS-59) is 
remarkable for the number of features, elaborate 
rock art, and unusual height (nearly 2 m). 
Among the rooms larger than 8 m 3 , there is a 



clear mode (and median) of 9-10 m 3 : 5 of the 13 
fall into that size range, and 8 fall between 9 m 
and 15 m 3 (see Figure 4.1). 

Using a t-test to compare the 29 
relatively complete Tsankawi chambers 
designated as habitation rooms with the 44 from 
Frijoles, the two groups are significantly 
different (means of 6.28 and 4.26 m 3 , t= -2.59, 
p =0.014). A similar comparison of rooms 
designated as storage rooms, however, finds no 
significant difference between groups of 19 and 
23 rooms (means of Tsankawi [1.42 m 3 ] and 
Frijoles [1.34 m 3 ], t= -0.24, p=0.811). 

When the chambers greater than 8 m 3 
are removed from the calculation, the Tsankawi 
mean becomes 2.7095 m 3 , which is smaller than 
the mean for all other groups but Group I. 
Perhaps large chambers were used for most 
functions, so that fewer activities were carried 
out in the remaining rooms, and they could be 
smaller. Alternatively, with access to larger 
rooms for habitation, the residents could devote 
more of the smaller rooms to storage. 

Assigning functions to chambers is a 
matter of speculation, but we based the 



FEATURES 113 

Table 4.3. Chamber Occurrence and Dimensions. 

A. Chamber Occurrence by Shape and Group 

Shape Group A Group F Group I Group M Tsankawi Total 



Oval 








2 





2 


4 


with hemispherical top 





1 











1 


Rectangular 


3 


1 


4 


8 


8 


24 


with hemispherical top 











1 





1 



Cylindrical 



Hemispherical 





4 


2 


19 


33 


58 


with rectangular top 














1 


1 


with pyramidal top 


1 














1 


Truncated cone 











2 


2 


4 


Truncated pyramid 


17 


11 


11 


17 


21 


77 


with hemispherical top 














2 


2 


with conical top 


1 














1 


with 2nd truncated 














1 


1 


pyramid 















Spherical 10 3 4 

Irregular 2 2 

Total 23 17 21 47 74 182 



114 CAVATE STRUCTURES 



Table 4.3. (continued) 



B. Mean Chamber Volumes in Cubic Meters 



Grouping 



Coefficient 
Standard of Variation 

Mean Deviation Minimum Maximum (CV) 



By shape 














Rectangular 


12 


3.86317 


2.34067 


0.7358 


9.1728 


60.6 


Hemispherical 


37 


3.72179 


3.99955 


0.3445 


15.4956 


107.5 


Truncated 
Pyramid 


69 


4.24722 


3.22900 


0.3512 


18.6326 


76.0 


By assigned 
function 














Habitation 


73 


5.04244 


2.69408 


0.5906 


15.1495 


53.4 


Storage 


42 


1.37660 


1.04885 


0.3445 


5.5545 


76.2 


Kiva 


7 


8.98803 


6.74858 


2.7211 


18.6326 


75.1 


By group 














Group A 


18 


3.60704 


2.45064 


0.5576 


7.5840 


67.9 


Group F 


13 


3.32241 


1.87310 


0.3512 


6.2586 


56.4 


Group I 


14 


2.85180 


1.35703 


0.7358 


5.5313 


47.6 


Group M 


29 


3.14225 


2.19974 


0.3445 


7.3764 


70.0 


Tsankawi 


57 


4.78904 


4.41563 


0.4523 


18.6326 


91.0 


By group and 
function 














Group A 














Habitation 


12 


4.92763 


1.83442 


2.2717 


7.5840 


37.3 


Storage 


5 


0.69419 


0.14392 


0.5576 


0.9307 


20.7 


Group F 














Habitation 


10 


4.11853 


1.25625 


2.2362 


6.2586 


30.5 


Storage 


3 


0.66867 




0.3512 


1.2032 





FEATURES 



115 



Table 4.3. (continued) 









Standard 






Coefficient 
of Variation 


Grouping 


n 


Mean 


Deviation 


Minimum 


Maximum 


(CV) 


Group I 














Habitation 


7 


3.18887 


1.34459 


1.5239 


5.5313 


42.2 


Storage 


5 


1.96018 


0.99595 


0.7358 


3.0620 


50.8 


Kiva 


1 


2.86365 










Group M 














Habitation 


15 


4.31373 


2.03163 


0.5906 


7.3764 


47.1 


Storage 


10 


1.55611 


1.57855 


0.3445 


5.5545 


101.4 


Kiva 


2 


2.88122 




2.7212 


3.0413 




Tsankawi 














Group 














Habitation 


29 


6.27539 


3.40041 


1.3315 


15.1495 


54.2 


Storage 


21 


1.41990 


0.82667 


0.5422 


3.2121 


58.2 


Kiva 


4 


13.05380 


6.22100 


4.8256 


18.6326 


47.7 



Grouping 



n 



C. Mean Heights in Meters 



Mean 



Coefficient 
Standard of Variation 

Deviation Minimum Maximum (CV) 



By assigned 
function 














Habitation 


70 


1.58 


0.441 


0.95 


3.93 


28.0 


Storage 


38 


1.11 


0.242 


0.75 


2.05 


21.7 


Kiva 


7 


1.62 


0.232 


1.25 


1.94 


14.3 



116 CAV ATE STRUCTURES 



Table 4.3. (continued) 



Grouping 


n 


Mean 


Standard 
Deviation 


Minimum 


Maximum 


Coefficient 

of Variation 

(CV) 


By group 














Group A 


16 


1.42 


0.271 


0.85 


1.81 


19.1 


Group F 


12 


1.39 


0.241 


0.97 


1.74 


17.3 


Group I 


14 


1.48 


0.411 


0.82 


2.05 


27.8 


Group M 


28 


1.32 


0.351 


0.69 


2.12 


26.6 


Tsankawi 


51 


1.47 


0.539 


0.75 


3.93 


36.6 



Note: Chambers with indeterminate function assigned are not shown. Table includes only chambers 
judged to be more than 70% complete. Heights have been corrected by adding fill depths. Volumes 
have been given for all shapes, heights for simple regular shapes. 



assignments on criteria including numbers and 
types of features, plastering, and size (Figure 
4.2). Although the "kiva" features are present 
in larger chambers at Tsankawi, those at the 
Frijoles groups are within the range of habitation 
room sizes recorded here (Figures 4.2, 4.3a). 
Rooms called storage rooms tend to be smaller, 
but there are some large examples as well 
(Figure 4.3b). Clearly, the distribution for 
"habitation" rooms shows a great deal of overlap 
with the other two categories. Most of the 
rooms to which no function was assigned are 
smaller (less than 3 m 2 ), though there are three 
larger ones as well (Figure 4.2). 

Hyland's (1986:101) results from Garcia 
Canyon show chamber volumes and heights 
considerably greater than those we found. He 
gives a mean volume of 9.62 m 3 with a range of 
0.99-24.42 m 3 , and of 25 measured cases his 
data contain 9 chambers greater than 12 m 3 . 
This disparity is partly a result of his having 
calculated volumes from the product of the 
maximum dimensions of the cavates involved, 
without taking into account how cavate rooms 
constrict toward the ceiling or how rarely they 
approximate rectilinear solids. If we calculated 



the volume for the largest chamber at Tsankawi 
following Hyland's procedure, the result would 
be more than 25 m 3 , larger than any of the 
volumes Hyland found. Still, the apparent 
tendency to build larger chambers at both 
Tsankawi and Hyland's more northern cavates 
may have some cultural significance. 

Exterior Door (Code 2) 

This feature type was used for all 
intentionally shaped openings to what is 
currently outside. Many such doors probably 
opened into masonry rooms when the sites were 
in use, but we assigned them to this type anyway 
(Table 4.4). 

Measurable doors are clearly mostly 
rectangular with some oval and trapezoidal 
variants (Figure 4.25); of the "circular" doors, 
three were considered full circles and one a half 
circle. Not surprisingly, 60 of the 65 recorded 
doors are found in the exterior wall of their 
rooms, with the other 5 occurring in exterior 
corners (2), the right wall (1), the top of the 
chamber (1), and inside another feature (1). 



FEATURES 117 




1 1 8 C AV ATE STRUCTURES 




FEATURES 



119 



R 



to 



CO 

c 
o 



52 

n 

E 

CO 

O 

(0 

> 



c 

CO 

c 
o 

• MB 

s 

■ ■■» 

-Q 
(0 



.2 s 

Sjff 

CO CO 5 S 

o o c c 

ILILhh 







(O 



© 00 CO * <M O 



r 



T 
M 



oo 






—r 



CM 



lunoo 



2! 



6 



I 






2 



fi 



•a 



S !! 

|| 

S3 
til 

?S 
« t 

i* 

3 Si 



i 



120 CAVATE STRUCTURES 



$ 



s 

E.2 



2 
o 

n 

E 

CO 



<D 
O) 

5 
O 

35 




CM 



o 

CM 



IO 



- 1— 
ID 



Uinoo 









FEATURES 



121 



Table 4.4. Exterior Door Occurrence and Dimensions. 

A. Exterior Door Occurrence by Shape and Group 



Shape 


Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


Rectangular 


6 


1 


2 


1 


12 


22 


Bowed rectangular 


2 


1 


2 


2 


5 


12 


Oval 





3 





2 


7 


12 


Trapezoidal 


1 


4 


1 





3 


9 


Circular portion 


1 











3 


4 


Triangular 





1 


1 


1 


2 


5 


Irregular 





1 











1 


Total 


10 


11 


6 


6 


32 


65 



Dimension 



B. Mean Dimensions of Oblong Doors in Meters 



n 



Mean 



Standard 
Deviation 



Minimum Maximum 



CV 



Width 


48 


0.68 


0.275 


0.30 


1.35 


40.4 


Height 


49 


1.00 


0.312 


0.48 


2.10 


31.1 


Wall Thickness 


44 


0.34 


0.157 


0.05 


0.88 


46.5 


Area 


48 


0.61 


0.334 


0.15 


1.98 


54.5 



Note: Only features considered to be 70% or more complete are included. 

B, coefficients of variation (CV) in all tables are calculated with unrounded means and 
standard deviations. These CV's will be different from those obtained using the rounded 
values in the tables, particularly when the values are very small. 

"Oblong includes rectilinear, oval, trapezoidal, and triangular shapes. 



122 CAVATE STRUCTURES 



Also predictably, all but one are located in 
cavate features. 

Even though some extreme cases are 
included, the most consistent dimension is the 
height, with a mean of exactly 1 m. The mean 
door width is two-thirds of the height, giving an 
indication of the basic shape of relatively 
complete doors. Some doors show evidence of 
having had lintels, which were probably stone 
(Figures 2.12 and 2.16 show doors with lintel 
grooves). In some cases the opening for the 
door may have been divided into a door and a 
smoke vent by a lintel. 

Exterior Opening (Code 26) 

This code was designed to record cavate 
openings that could not be called doors. The 
absence of a formal door has two main causes: 
loss of sections of cliff, which would have been 
weakened by the placement of both room and 
door, and, especially in Frijoles, loss of 
masonry chamber closings. The exterior 
opening, then, is a cross-section of the room at 
the cliff and as such is a potentially useful 
source of information about the room. Since 
completely enclosed cavates are the exception 
rather than the rule, this should be a very 
commonly used code. It seems, however, that 
its intent was somewhat misunderstood by the 
recorders, who probably did not record it in all 
cases. Once again, deciding whether a shallow 
feature has an "exterior opening" or is merely a 
"back wall" requires the exercise of judgment. 

All but three of these features are 
recorded as being in the exterior wall (the three 
exceptions are in the left and right walls). The 
majority are semicircular, and the seven 
trapezoidal cases corroborate the tendency of 
cavates to constrict toward the ceiling (Table 
4.5). Since a large number of noncavates are 
back walls, noncavates are unlikely to have 
either doors or exterior openings. In order to 
examine the variable occurrence of openings at 
the various groups, Table 4.6 compares numbers 



of cavates, doors, exterior openings and 
noncavates. Since only cavates are likely to 
have doors or exterior openings, the total 
number of openings should be roughly equal to 
the number of cavates present. At Groups A 
and F the large number of non-cavates accounts 
for the relatively small number of openings 
recorded. Over half of the exterior openings 
recorded are at Tsankawi, though Tsankawi 
accounts for around a third of the total numbers 
of rooms recorded. The number of doors and 
openings recorded at Tsankawi accounts well for 
the cavates recorded there, though several 
Tsankawi rooms do have more than one 
opening. The under-representations, then, are at 
Groups M and I; this results in part from a lack 
of full understanding of the use (and usefulness) 
of this category in the early recording at Group 
M. Group M, in fact, probably has more 
openings that would have been profitably 
recorded in this category than the other groups. 
I am unable to account for the low count at 
Group I. 

Interior Door (Code 19) 

Doors leading from one cavate chamber 
into another, usually providing access to a 
smaller inner chamber. 

As is true for the exterior doors, the 
most consistent dimension for interior doors is 
the height. On average, interior doors are 
considerably lower than exterior doors (0.7 m 
versus 1 m) and somewhat narrower (0.56 m 
versus 0.68 m). As can be seen from the 
dimensions, some interior doors are so small 
that a modern adult finds it difficult to pass 
through them (Table 4.7). 

Natural Wall (Code 17) 

The great majority of remaining walls in 
the cavates recorded are entirely of "living" 
rock, without exception tuff. Even though they 
are carved from the cliff, distinct walls are 
usually definable, though this attribute varies 



FEATURES 



123 



Table 4.5. Exterior Opening Occurrence and Dimensions. 

A. Exterior Opening Occurrence by Shape and Group 



Shape 


Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


Rectangular 


2 








1 


6 


9 


Bowed rectangular 











1 


4 


5 


Oval 


4 











11 


15 


Trapezoidal 


1 


1 





3 


2 


7 


Circular portions 


8 


3 


1 


2 


14 


28 


Triangular 





1 


1 





2 


4 


Irregular 


1 








3 





4 


Total 


16 


5 


2 


10 


39 


72 



B. Mean Dimensions by Shape in Meters 



Dimension 



n 



Mean 



Standard 
Deviation 



Minimum Maximum 



CV 



Oval 














Width 


12 


1.98 


0.958 


0.80 


3.00 


48.4 


Height 


12 


1.01 


0.367 


0.96 


1.85 


25.5 


Area 


12 


0.849 


0.584 


0.204 


2.227 


68.8 


Rectangular 














Width 


14 


1.13 


0.706 


0.20 


2.60 


62.6 


Height 


14 


1.00 


0.250 


0.55 


1.49 


25.0 


Area 


14 


1.125 


0.862 


0.088 


3.013 


76.6 


Round 














Diameter 


8 


0.89 


0.40 


0.30 


1.60 


44.7 


Area 


8 


0.651 


0.503 


0.071 


1.609 


77.4 


Semicircular 














Width 


9 


1.86 


0.756 


0.85 


3.00 


40.6 


Height 


6 


1.02 


0.302 


0.70 


1.55 


29.5 


Area 


9 


1.657 


1.156 


0.425 


3.535 


69.7 



124 CAVATE STRUCTURES 



Table 4.6. Occurrence of Doors and Openings by Group and Cavate Type. 

Group A Group F Group I Group M Tsankawi Total 



Openings 














a) Exterior Doors 


10 


11 


6 


6 


32 


65 


b) Exterior Openings 


16 


5 


2 


10 


39 


72 


c) Total Openings 
(a+b) 


26 


16 


8 


16 


71 


137 


Rooms 














d) Cavates + 
Noncavates 


73 


60 


37 


65 


118 


353 


e) Noncavates 


48 


45 


16 


19 


56 


184 


Co-occurrence 














f) Openings + 
Noncavates (c+e) 


74 


61 


24 


35 


127 


321 


Difference (f-d) 


+ 1 


+ 1 


-13 


-30 


+9 


-32 



FEATURES 



125 



Table 4. 7. Interior Door Occurrence and Dimensions. 



Shape 



A. Interior Door Occurrence by Shape and Group 

Group A Group F Group I Group M Tsankawi 



Total 



Rectangular 
Oval 

Trapezoidal 
Round 
Triangular 
Total 




3 

1 



4 












3 

2 
2 

1 
8 



2 
5 


1 
8 



1 
9 

8 

1 
19 



6 
19 

3 

8 

3 

39 



B. Chamber Location by Shape 











Exterior 










Right 






Wall& 


Back 






Shape 


Wall 


Back Wall 


Left Wall 


Corners 


Corners 


Other 


Total 


Rectangular 


1 


1 


2 


2 








6 


Oval 


5 


3 


4 


4 


2 


1 


19 


Trapezoidal 


1 








2 








3 


Round 


1 


1 


3 


2 


1 





8 


Triangular 


1 














2 


3 


Total 


9 


5 


9 


10 


3 


3 


39 



Dimension 



C. Mean Dimensions ofNonround Doors in Meters 



Mean 



Standard 
Deviation 



Minimum 



Maximum 



CV 



Width 


22 


0.56 


0.227 


0.30 


1.35 


40.1 


Height 


22 


0.70 


0.176 


0.41 


1.10 


25.0 


Wall Thickness 


10 


0.28 


0.136 


0.08 


0.50 


49.1 


Area 


22 


0.338 


0.1814 


0.105 


0.672 


53.7 



Note: C, only doors thought to be 70 percent or more complete are included. Nonround includes 
rectangular, oval, trapezoidal, and triangular shapes. 



126 CAVATE STRUCTURES 



from rooms with four squared corners to nearly 
spherical small chambers. In the latter cases, 
definition of discrete walls is of course more 
subjective. 

The recorded shapes of natural walls do 
show some patterning (Table 4.8). About three- 
quarters of the walls at each group are 
trapezoidal, reflecting the preference for 
chambers in the shape of truncated, four-sided 
pyramids. The Group A cavate sample contains 
the lowest percentage of trapezoidal walls, the 
difference being made up by higher percentages 
of rectangular, semicircular, and triangular 
walls. The common shapes (rectangles and 
trapezoids) occur in very similar proportions, 
while the odd shapes, such as triangles, 
semicircles, and lines, occur much more often in 
walls other than the back wall. This is because 
walls other than the back wall are much more 
likely to be partial than are back walls. 

Wall heights are clearly less variable 
than widths. The right, back, and left walls all 
fall into similar ranges, with each having some 
extreme values less than 50 cm. The means are 
strikingly close to 1.2 m for exterior, left, and 
right walls, while the back walls average closer 
to 1.5 m. This difference doubtless results from 
the fact that full heights for back walls of 
noncavate rooms are generally larger than for 
rooms excavated from the tuff. Predictably, the 
areas vary the most, as the variability of both 
linear dimensions is compounded. 

Plastering 

As can be seen in the series of plaster 
color, coat, and height tables, the average height 
of plaster on all types of reasonably intact walls 
(those with FF greater than 0.7) that show some 
plastering and have suffered moderate or no 
natural damage is 0.86 m, with a range of 0.13- 
1.77 m (Tables 4.9-4.11). This value is quite 
consistent from wall to wall and from group to 
group, though the mean given is subject to at 
least two complicating factors. On one hand the 



minimum value of 13 cm probably does not 
represent a full plaster height, and plaster loss is 
likely to have reduced measurements in some 
cavates. On the other hand some chambers have 
a single coat of plaster extending to the ceiling, 
while other coats stop short of it; in such cases, 
the plaster height was measured to the height of 
the coat that goes farthest up the wall, which has 
probably caused some inflation in the height 
values. This single coat, like the single coats 
often seen in storage rooms, seems to be a 
rough, scratch coat. Heavily or long used 
habitation rooms can have numerous coats of 
smooth plaster. Counting visible coats, and 
without dissecting the plaster, the largest number 
of coats we observed was 10. Carlson and 
Kohler (1989:53) report 12 coats in Group M, 
but some of those may relate to the construction 
of a niche. 

Although some walls are plastered to the 
ceiling, the majority are plastered only to a level 
well below the ceiling (Figures 3.3, 4.12). 
Small wall holes are more abundant near the top 
of the plaster. The largest mean is for back 
walls, perhaps due to preservation and to the 
presence of some higher values from masonry- 
fronted rooms. In reading the plaster tables one 
must recall that the counts are taken from walls 
rather than chambers. Chambers with multiple 
coats of plaster show both smoked and 
unsmoked coats, suggesting variation or changes 
in chamber function, or differences in the 
periodicity of plastering. The low frequency of 
walls with preserved plaster at Tsankawi is 
clearly evident. 

Of 524 natural and masonry walls 
recorded in cavates, 339 (65 percent) had 
enough plaster remaining to be recorded. 
Plaster colors are listed in Table 4.11, which 
shows that by far the most common color was 
tan. The second most common category, walls 
with multiple colors of plaster, results from the 
number of walls with more than one coat of 
plaster visible and from variation in color across 
the wall. 



FEATURES 



127 



Table 4.8. Natural Wall Occurrence and Dimensions. 

A. Natural Wall Occurrence by Shape and Group 



Shape 


Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


Rectangular 


34 


12 


18 


23 


35 


122 


Bowed rectangular 





5 





2 


2 


9 


Oval 








1 


1 





2 


Trapezoidal 


88 


75 


66 


92 


218 


539 


Circular portions 


1 


7 





2 


2 


12 


Triangular 


13 


5 


4 


2 


8 


32 


Hemispherical 


1 














1 


Linear 











1 





1 


Total 


137 


104 


89 


123 


265 


718 



B. Chamber Location by Shape 



Shape 


Right Wall 


Back Wall 


Left Wall 


Exterior Wall 


Total 


Rectangular 


19 


77 


19 


7 


122 


Bowed rectangular 


2 


5 


2 





9 


Oval 





2 








2 


Trapezoidal 


137 


205 


131 


65 


538 


Circular portions 


2 


7 


2 


1 


12 


Triangular 


12 


5 


9 


6 


32 


Hemispherical 





1 








1 


Linear 








1 





1 


Total 


172 


302 


164 


79 


718 



128 



CAVATE STRUCTURES 



Table 4.8. (continued) 



C. Mean Dimensions by Location in Meters 



Wall 



Mean 



Standard 
Deviation 



Minimum Maximum 



CV 



Width 














Right 


102 


1.61 


0.981 


0.40 


7.45 


60.9 


Back 


236 


2.18 


0.858 


0.44 


5.00 


39.3 


Left 


100 


1.51 


0.708 


0.37 


5.30 


47.0 


Exterior 


44 


2.03 


1.265 


0.30 


8.30 


62.4 


Height 














Right 


102 


1.21 


0.374 


0.38 


2.25 


30.9 


Back 


234 


1.51 


0.467 


0.35 


3.05 


31.0 


Left 


100 


1.22 


0.365 


0.55 


2.40 


30.0 


Exterior 


44 


1.23 


0.382 


0.37 


2.10 


31.1 


Area 














Right 


101 


1.77 


1.275 


0.14 


7.25 


72.2 


Back 


235 


3.08 


1.885 


0.22 


10.23 


61.2 


Left 


100 


1.67 


1.078 


0.17 


7.58 


64.5 


Exterior 


44 


2.32 


1.943 


0.16 


11.57 


83.9 



FEATURES 129 



Table 4.9. Plaster Coats by Group, Wall, and Function. 

A. Plaster Coats by Group 



Coats 


Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 





8 


2 


2 


1 


28 


41 


1 


5 


5 


12 


33 


16 


71 


2 


5 


4 


3 


10 


14 


36 


3 


8 


7 


1 


6 


6 


28 


4 


7 


1 





4 


4 


16 


5 


4 


3 


2 


2 


1 


12 


6 





2 


1 


1 





5 


7 


3 


1 





1 





5 


8 


1 














1 


9 





1 











1 


10 


2 














2 


Total 


43 


26 


21 


58 


69 


218 



B. Plaster Coats by Chamber Wall 

Coats Right Wall Back Wall Left Wall Exterior Wall Total 



1 

2 

3 

4 

5 

6 

7 

8 

9 

10 

Total 



14 


11 


12 


4 


41 


19 


20 


21 


11 


71 


14 


12 


9 


1 


36 


7 


12 


7 


2 


28 


3 


5 


5 


3 


16 


4 


4 


4 





12 


1 


4 








5 





4 


1 





5 





1 








1 





1 








1 


1 








1 


2 


63 


74 


59 


22 


218 



130 CAVATE STRUCTURES 



Table 4.9. (continued) 



C. Wall Plaster Coats by Assigned Function 

Coats Habitation Storage Kiva Total 

16 

1 23 

2 25 

3 23 

4 11 

5 7 

6 5 

7 4 

8 1 

9 1 
10 2 

Total U8 71 24 213 

Note: A-C, only walls considered to be more than 70 percent present and to have had slight or no natural 
damage are included. C, five cases with unassigned function are not shown. 



25 





41 


40 


3 


66 


6 


5 


36 





5 


28 





5 


16 





5 


12 








5 





1 


5 








1 








1 








2 



FEATURES 



131 



Table 4.10. Mean Plaster Height and Number of Coats by Group and Wall. 



Height and Coats 


n 


All walls height (m) 


151 


All walls coat (n) 


151 


Group 




A height 


35 


A coats 


35 


F height 


21 


F coats 


21 


I height 


15 


I coats 


15 


M height 


52 


M coats 


52 


Tsankawi height 


28 


Tsankawi coats 


28 


Wall 
Right wall height 


41 


Right wall coats 


41 


Back wall height 


54 


Back wall coats 


54 


Left wall height 


41 


Left wall coats 


41 


Exterior wall height 


15 


Exterior wall coats 


15 





Standard 








jan 


Deviation 


Minimum 


Maximum 


CV 


0.86 


0.304 


0.13 


1.77 


35.3 


2.7 


1.90 


1 


10 


70.6 


0.90 


0.210 


0.24 


1.34 


23.2 


3.9 


2.37 


1 


10 


61.1 


0.85 


0.279 


0.55 


1.77 


32.6 


3.0 


1.91 


1 


9 


64.7 


0.68 


0.323 


0.13 


1.21 


47.7 


2.2 


1.74 


1 


6 


79.1 


0.93 


0.332 


0.27 


1.68 


35.7 


2.1 


1.55 


1 


7 


73.5 


0.78 


0.319 


0.23 


1.52 


40.63 


2.3 


1.12 


1 


5 


48.4 



0.86 


0.333 


0.24 


1.52 


38.8 


2.5 


1.76 


1 


10 


71.5 


0.93 


0.263 


0.27 


1.68 


28.3 


3.2 


2.12 


1 


9 


66.6 


0.78 


0.257 


0.23 


1.26 


32.7 


2.3 


1.37 


1 


9 


59.0 


0.84 


0.435 


0.13 


1.77 


52.1 


2.5 


1.28 


1 


10 


66.2 



Note: Only plastered walls considered to be more than 70 percent present and to have had slight or no 
natural damage are included. Only walls with plaster height greater than 0.10 m included. 



132 CAVATE STRUCTURES 



Table 4.11. Plaster Color by Group. 



Color 



Group A Group F Group I Group M Tsankawi 



Total 



Tan 


38 


17 


34 


60 


38 


187 


Brownish 


2 


6 





3 


14 


25 


Reddish 


1 














1 


Black 





1 





4 





5 


White to gray 














3 


3 


Several colors 


22 


18 


22 


40 


16 


118 



Total 



63 



42 



56 



107 



71 



339 



Note: Shows all cavate walls with plaster recorded. 



Masonry Wall (Code 18) 

We found few walls constructed entirely 
of masonry in the sample recorded in 1986. We 
observed some such walls in unrecorded groups, 
including upper Group F and Group B. 
Although there can be little doubt that they were 
once common at these various sites, preserved 
portions are not visible without excavation. 

The very low frequency of remaining 
masonry walls is quite apparent (Table 4.12). 
Those recorded are mostly small remnants. The 
use of masonry for closing and separating rooms 
is visible in the chamber locations of masonry 
walls. We found no masonry walls in the 
section of Group F we recorded, but observed 
several in the upper part of the group. The 
presence of masonry walls at Group I is quite 
out of proportion to the size of that group, as is 
the absence of masonry at Tsankawi. Tsankawi 
is both the most heavily visited group as well as 
the group with the greatest relative use of wholly 
excavated chambers. Masonry was surely 
present in the Tsankawi group but is now 
missing. Exterior masonry is also absent at 
Group M; the "walls" in that group are two 



sides of the very well-preserved interior divider 
between M-59 and M-60. 

Masonry and Tuff Wall (Code 43) 

We recorded three examples of these 
walls, all at Group A. They are located at the 
exterior, the right, and the left of their 
respective rooms. The low count of this type of 
feature is unquestionably due to the low 
preservation rate of masonry at all five study 
groups. Tsankawi has several examples of what 
appear to be low tuff wall bases, which almost 
certainly had masonry upper walls. On the 
whole, however, walls were probably either all 
natural or all masonry. 

Chamber Corner (Code 39) 

Corners were recorded primarily as 
means of triangulating the location of the 
compass used to take azimuths for the features 
so measured. Although we recorded direction 
for each feature that could be located as a point, 
we took distances from the compass to the 
feature only for triangulation points. 



FEATURES 



133 



Table 4.12. Masonry Wall Occurrence and Dimensions. 

A. Masonry Wall Occurrence by Shape and Group 



Shape 



Shape 



Group A Group F Group I Group M Tsankawi 



B. Chamber Location by Shape 

Right Wall Back Wall Left Wall Exterior Wall 



Total 



Rectangular 








3 


2 





5 


Trapezoidal 


1 














1 


Total 


1 





3 


2 





6 



Total 



Rectangular 


2 





1 


2 


5 


Trapezoidal 











1 


1 


Total 


2 





1 


3 


6 



C. Mean Area by Shape in Meters 



Shape 


n 


Mean 


Standard 
Deviation 


Minimum 


Maximum 


CV 


Rectangular 
Trapezoidal 


5 
1 


1.375 
4.410 


1.146 


0.141 


11.567 


71.7 



134 CAVATE STRUCTURES 



Floor (Code 3) 

The floors of the rooms recorded are 
frequently not visible, either because of fill or 
because they are missing. Where possible 
without excavation or destruction, we recorded 
coats of plaster and dimensions. The floor and 
the other enclosing parts of chambers are both 
features and locations in this recording system, 
which leads to some redundancy in their coding 
(i.e., code: feature = floor, code: part=floor). 

The frequencies of floors recorded show, 
among other things, the visibility of floors at the 
various groups (Table 4.13). Thus, Group I and 
Group F have low counts because the former has 
many filled chambers and some partial cavates, 
while the latter has more noncavates. The very 
low count at Tsankawi, however, seems an 
underrepresentation, possibly caused by 
development of a casual attitude toward 
recording floors as features as the field season 
progressed. It is also true, however, that the 
Tsankawi group has many chambers containing 
substantial fill. Means for width and length are 
the same, indicating the tendency of cavates 
toward square or round bases, rather than 
toward rectangular, as would be more likely for 
masonry rooms. 

Replastering of walls and floors gives a 
crude index of intensity and duration of room 
use, though the interval between pi aster ings is 
unknown and probably varied, and it is possible 
that at least two coats-one scratch and one 
finish-may have been minimal. The floor 
replastering data show that a few features at 
Group A were probably quite heavily used, since 
they have up to seven coats of plaster (Table 
4.14). More than half of the observable cases 
show two or three coats of floor plaster. For 
rooms thought likely to have been storage 
rooms, the floor plastering data are quite 
different from the wall plastering data: while 
storage room walls have no plaster or only one 
coat, five of the nine "storage room" floors 
recorded have more than two plaster coats. 



Ceiling (Code 27) 

A major reason for including ceilings as 
features was to record the presence or absence 
of smoking; a separate entry also allowed for 
recording specific notes about the ceiling. 
Generally, we did not take measurements for 
ceilings since they are defined by the tops of 
walls, which we did measure. 

Of the 122 ceilings recorded, only 14 
are recorded as having plaster, and of those only 
2 have two coats. The tables for smoking show 
that most cavate ceilings are smoked (Table 
4.15). The unplastered and smoked ceilings 
include the majority of smaller rooms, which we 
thought most likely to have served for storage. 
The friability of the tuff and the absence of 
plaster would probably mean a steady sifting of 
fine particles on the occupants and contents of 
cavate chambers. Steen (1977:17; 1979) has 
suggested that cavates were intentionally smoke- 
blackened as part of chamber preparation, and 
smoke blackening in storage rooms supports that 
argument. Chamber interiors do appear to be 
somewhat more resistant to erosion than the 
immediately surrounding tuff, since a sort of 
rind is sometimes visible at the front edge of 
cavates. Heiken even suggests that cavate walls 
might have been prepared by heating and then 
dousing, which would accelerate formation of a 
patina on the tuff (G. Heiken, personal 
communication, 1986; see also Heiken 1979). 
G. White (1904:67) made an interesting 
observation about the tuff in Cappadocia: "This 
rock is so soft that it can be slowly whittled 
away with a knife, and doors, windows, 
stairs . . . and rooms greater and smaller are 
easily worked in it, though it does not wear 
away rapidly under natural agencies, and its 
surface hardens on exposure to the air." 

There are reasons to remain somewhat 
skeptical of the attribution of most smoking in 
chambers to intentional preparation rather than 
to use. The amount of smoking that remains in 
cavates is quite variable, and large chambers 



FEATURES 135 



Table 4.13. Floor Occurrence and Dimensions. 

A. Floor Occurrence by Shape and Group 



Shape 


Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


Rectangular 


10 


3 


5 


6 


2 


26 


Bowed rectangular 


1 








2 





3 


Trapezoidal 


2 








8 





10 


Circular portion 


1 








8 


3 


12 


Triangular 


2 








5 


1 


8 


Linear 











1 





1 


Irregular 











1 





1 


Total 


16 


3 


5 


31 


6 


61 



Function 



B. Mean Floor Areas for All Shapes in Meters 



Mean 



Standard 
Deviation 



Minimum Maximum 



CV 



All rooms 


25 


2.40 


1.36 


0.47 


4.86 


56.7 


Habitation 


16 


2.97 


1.17 


0.92 


4.86 


39.4 


Storage 


6 


0.93 


0.39 


0.47 


1.60 


41.9 


Kiva 


2 


3.13 




2.80 


3.46 





Note: B, includes only rooms with 15 cm or less of fill and floors more than 70 percent complete. 



2 








4 


4 


10 


3 


1 


1 


7 





12 


4 





4 


9 


1 


18 


4 


2 





10 


1 


17 











1 





1 


1 














1 


1 














1 


1 














1 



136 CAVATE STRUCTURES 

Table 4.14. Floor Plaster Coats by Group and Function. 

A. Floor Plaster Coats by Group 

Coats Group A Group F Group I Group M Tsankawi Total 


1 
2 
3 
4 
5 
6 
7 



Total 16 3 5 31 6 61_ 

B. Floor Plaster Coats by Assigned Function 
Coats Habitation Storage Kiva Total 



1 
2 
3 
4 
5 
6 
7 

Total 39 9 6 54_ 

Note: B, seven cases with unknown function not shown (5 of these have no coats). 



3 


2 





5 


8 


2 


1 


11 


12 


4 


2 


18 


13 


1 


2 


16 








1 


1 


1 








1 


1 








1 


1 








1 



Table 4.15. Ceiling Occurrence and Smoking by Group and Function. 

A. Occurrence by Shape and Group 



FEATURES 137 



Shape 


Group 


A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


Rectangular 


10 




8 


9 


10 




22 


59 


Oval 







1 










1 


2 


Trapezoidal 


3 







3 


6 




6 


18 


Circular portion 


3 










5 




13 


21 


Triangular 


3 







1 


3 




3 


10 


Cylindrical 







3 


1 


3 




2 


9 


Total 


19 




12 


14 


27 




47 


119 








B. Ceiling Smoking by Group 








Absent/Present 


Group 


A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


Smoking absent 


3 







4 


6 




7 


20 


Smoking present 


17 




12 


10 


22 




41 


102 


Total 


20 




12 


14 


28 




48 


122 




C. 


Ceiling Smoking by Assigned Function 








Absent/Present 


Habitation 


Storage 




Kiva 




Total 


Smoking absent 






8 


11 









19 


Smoking present 






68 


22 




8 




98 


Total 






76 


33 




8 




117 



138 CAVATE STRUCTURES 



with many features seem to have the thickest 
carbon deposits. Moreover, successive coats of 
wall plaster can be seen to have become quite 
black. A photograph of a Cappadocian kitchen 
cavate (Blair 1970:129) shows extremely black 
walls, presumably resulting from cooking. Tuff 
is a porous material, and it seems likely that 
smoke black would readily adhere to it. 
Whether smoking was intentional or not, it 
probably helped stabilize cavate ceilings. 
Steen's and Heiken's suggestion of intentional 
smoking and heating of cavates immediately 
after excavation could be tested by careful 
removal of a section of wall and floor plaster to 
ascertain whether or not burning took place 
before any plastering. 

Passage (Code 42) 

A few of the cavate chambers studied 
were connected to other chambers by short 
tunnels that could not be considered rooms on 
their own. This means of reaching another 
room was less frequent than simple doorways. 
Instances of this feature were found at Group A 
(between A-32 and A-34), Group M (connecting 
M-56 and M-58) and Tsankawi (between TS-52 
and TS-53). All three connect rooms with floors 
at somewhat different levels, so that the passages 
slope. The Tsankawi example is quite short, but 
it is more than just an interior door. 

Combined Chamber (Code 53) 

The procedure of recording upper and 
lower chambers separately because of differing 
shapes means that volumes of these chambers 
are not comparable to those recorded as single 
shapes. After computing volumes for the 
constituent parts, we combined them and entered 
them under this separate feature type with a part 
code for chamber, enabling them to be included 
in comparisons and averages. 

Compass Location Point (Code 51) 

Where possible, compass locations were 



tied to discrete, identifiable features. In some 
cases this was not possible. The points selected 
are described in the notes to the feature, as is 
the distance from the compass to the feature. 
Feature azimuths and heights in combination 
with the triangulated compass location allow 
reidentification of features. 

Floor Features 

Firepit (Code 4) 

This code was reserved for formally 
constructed, lined pits that showed evidence of 
having been burned. We found relatively few 
because they are quite consistently placed at the 
outer edge of the chamber and thus are subject 
to deterioration. In addition, we did virtually no 
excavation, and depressed floor features are 
usually filled even in rooms containing very 
little floor fill. Especially given the high 
frequency of smoked chambers and of smoke 
holes, there is little doubt that firepits are 
substantially underrepresented relative to their 
prehistoric frequency (Table 4.16). 

Visible firepits are especially abundant at 
Group A. We observed intact examples in 
unrecorded portions of Groups F and M. The 
recorded features vary considerably in size, 
which is surprising since they are generally 
located in small rooms. 

Floor Burn (Code 5) 

This feature covers cases in which there 
is a clearly localized burn on a floor, although 
no formal firepit was constructed. Though they 
are common in excavated rooms in Chaco, we 
observed relatively few in the cavates. We 
recorded a total of six floor burns, four in 
Group M and one each at Group A and 
Tsankawi. Group M thus held the largest 
number of floor burns as well as floors recorded 
(Table 4.13). 



FEATURES 



139 



Table 4.16. Firepit Occurrence and Dimensions. 

A. Firepit Occurrence by Shape and Group 



Shape 




Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


Rectangular 




2 








1 


1 


4 


Oval 




1 





1 





2 


4 


Trapezoidal 




1 














1 


Round 
















2 


2 


Rectangular 


solid 


8 


1 


2 


5 





16 


Total 




12 


1 


3 


6 


5 


27 



B. Mean Dimensions by Grouped Shape in Meters 









Standard 








Shape 


n 


Mean 


Deviation 


Minimum 


Maximum 


CV 


Oblong 














Length 


13 


0.56 


0.175 


0.30 


0.93 


31.5 


Width 


13 


0.44 


0.159 


0.23 


0.70 


36.1 


Volume 


13 


0.02381 


0.01682 


0.0000 


0.0499 


70.6 


Round diameter 


2 


0.59 




0.40 


0.78 





Note: Only features considered more than 70 percent present are included, oblong includes oval, 
rectangular, and rectangular solid. 



140 CAVATE STRUCTURES 



Subfloor Pit (Code 7) 

This feature category includes pits for 
which we inferred no specific function. It is, 
then, a catchall for features that do not fall into 
categories such as firepits, probable postholes, 
or pits for loom anchors. Again, since we did 
not excavate, we observed such floor features at 
a rate far less than their probable actual 
occurrence (Table 4.17). At Tsankawi, for 
example, the fill and floor of TS-59 were 
vandalized after the room was recorded, 
exposing several pits and floor features that were 
not visible at the time of recording. 

Though it was not possible to measure 
depths for some floor pits, they are all assumed 
to have depth. Thus, pits recorded as 
rectangular and as rectangular solids have been 
included in the same counts, as have round and 
cylindrical shapes. The volume means indicate 
how many of these features we were able to 
fully measure. The vandalized examples at 
Tsankawi suggest that some such features may 
be a good deal deeper than the 19 cm indicated 
as a maximum. As might be expected, there 
appear to be at least two categories here: those 
more like postholes (round in plan, with lower 
volumes) and those more like storage pits (oval 
or rectangular with higher volume). 

Floor Depression (Code 36) 

This is another descriptive code for 
features of unknown function. Some of these 
features are of appropriate size and shape to 
have been pot rests (see Figure 4.4). 

Just as for wall depressions, we recorded 
a disproportionate number of floor depressions 
for Group A. One room there (A-47) contains 
two plastered examples 19-20 cm in diameter 
that are especially reminiscent of pot rests 
(Figure 4.4). There are a couple of much larger 
features in this group, but the majority fit into a 
size range appropriate for pot rests (Table 4.18). 



Floor Pit Complex (Code 45) 

We observed this feature type only at 
Tsankawi. It consists of clustered groups of 
more or less regular, concave depressions in the 
tuff (Figure 4.5). The age of these pits was 
uncertain, and so was their function. Because 
they are in the tuff itself, it is at least possible 
that they would have been covered by plaster 
flooring, and they may have resulted from the 
process of excavating the chamber out of the 
tuff. Since they are found inside intact cavates 
with no other signs of severe weathering, they 
cannot have been produced by natural erosion. 
They are almost certainly manmade but may 
have been produced by modern visitors. Many 
of the individual pits bear some resemblance to 
axe-sharpening grooves, and some may have had 
a similar function. There are several possible 
reasons why this feature type was recorded only 
at Tsankawi: Tsankawi has more exposed tuff 
floors and receives heavier modern visitation. It 
is also possible that there was some cultural or 
functional basis for the difference. 

Posthole (Code 46) 

Postholes are another feature type 
observed only at Tsankawi. We recorded 9 as 
parts of cavates and noted 9-12 more as exterior 
to recorded rooms. Some of the recorded 
examples are near the edges of chambers and 
probably formed part of partitions or closing 
structures. Others are interior and may only 
resemble postholes in size and shape; they may 
have been used for loom uprights or some other 
function. The exterior postholes are located on 
the bedrock "terraces" in front of several groups 
of Tsankawi rooms (TS-15, TS-22-TS-25, TS- 
54-TS-59). In these areas they are likely to have 
been part of masonry (and jacal?) structures. 
The use of more wood and less stone seems 
likely given the relative scarcity of rubble in the 
areas outside the middle and upper Tsankawi 
cavates. None of these features suggests the use 
of large beams: the greatest recorded diameter is 



FEATURES 141 



Table 4.17. Floor Pit Occurrence and Dimensions. 

A. Floor Pit Occurrence by Type and Group 



Shape 


Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


Oval 











1 


4 


5 


Rectangular solid 


1 





1 


1 


1 


4 


Cylindrical 


2 








4 


2 


8 


Hemispherical 


1 





1 





2 


4 


Total 


4 





2 


6 


9 


21 



Shape 



B. Mean Dimensions by Shape in Meters 

Standard 
Mean Deviation Minimum Maximum CV 



All shapes 
volume 

Cylindrical 
volume 

Cylindrical 
diameter 

Cylindrical 
depth 



11 



0.01896 



0.01013 



0.26 



0.14 



0.02942 



0.01564 



0.181 



0.054 



0.00096 



0.09 



0.05 



0.10070 155.2 



0.00095 0.03802 154.4 



0.55 



0.19 



69.4 



37.7 



142 CAVATE STRUCTURES 




Figure 4.4. Group of three floor depressions 
in A-47. The depressions are in the floor 
plaster and are extremely suggestive of pot 
rests. The very symmetrical ledge in the 
wall behind (recorded as a wall niche) is 
somewhat unusual in shape, though a 
similar feature was observed in Group F as 
well. 



ridge seems to divide it into unequal portions, 
the lower side usually the smaller. These 
features may be related to mealing complexes, 
but we have little evidence and a fairly small 
sample. Hewett and others call them sleeping 
ridges. Future workers in cavates should be on 
the alert for complexes involving floor ridges, 
slots, metate rests, and wall depressions (Figures 
3.1, 4.6, 4.7). Again, Group M, which had the 
most floors visible, also had the preponderance 
of floor ridges: 7 of the 11 recorded. One floor 
ridge was recorded at Tsankawi and three at 
Group A. The notion that at least some floor 
ridges functioned to form partitions is supported 
by our observation that the crest of several of 
these ridges seems to be broken away, as would 
have happened with the removal of a mortared- 
in plank. 

Metate Rest (Code 40) 

No metates were observed during the 
1986 field season, either in place or on the 
slopes in front of the groups studied. Features 
called metate rests were reduced to inclined 
adobe lines on walls in all but one case; these 
lines extend from floor level to heights of 
around 20 cm (Figure 4.7a, Table 4.19). The 
lines are almost certainly the remnants of plaster 
cementing metates in place, but empirical proof 
for this statement is largely lacking. The 
exception is located at Group M and consists of 
an adobe ramp located a few centimeters away 
from the wall (Figure 4.6). 



20 cm, and the average is around 12 cm, with 
unexcavated depths ranging from 7 cm to 40 
cm. 

Floor Ridge (Code 6) 

Several rooms containing numerous 
features and having observable floors exhibit a 
raised floor feature, usually running across the 
width of the room (Figure 4.6). Sometimes the 
floor is lower on one side of the ridge than on 
the other. Rather than bisecting the room, the 



As can be seen in M-60 (Figure 4.7), 
mealing activities likely resulted in a complex of 
features: metate rests and worn areas in the wall 
from grinding, and perhaps slots and floor 
ridges for storage areas. To evaluate this 
possibility and to examine the occurrence of 
mealing features, Table 4.20 gives the 
frequencies of occurrence and co-occurrences of 
these feature types, in cavate rooms only. Given 
the invisibility of the majority of floors, the 
inventory is certainly partial. Only wall 
depressions less than 15 cm above the floor are 



FEATURES 143 



Table 4.18. Floor Depression Occurrence and Dimensions. 

A. Floor Depression Occurrence by Shape and Group 



Shape 


Group A 


Group F 


Group I 


Group 


M 


Tsankawi 


Total 


Rectangular 


2 
















2 


Oval 


1 








1 




1 


3 


Round 


2 








1 







3 


Hemispherical 


2 








1 




1 


4 


Total 


7 








3 




2 


12 



Dimension 



B. Mean Dimensions by Shape in Meters 



Mean 



Standard 
Deviation 



Minimum Maximum 



CV 



Diameter 


10 


0.24 


0.164 


0.05 


0.60 


69.0 


Depth 
(circular) 


10 


0.09 


0.106 


0.00 


0.30 


113.9 


Volume 


10 


0.00184 


0.00363 


0.00012 


0.01001 


197.1 


Rectangular 
length 


2 


0.22 




0.21 


0.22 




Rectangular 
width 


2 


0.22 




0.21 


0.22 





144 CAVATE STRUCTURES 




Figure 4.5. "Floor pit complex" in TS-55. The depressions are in the tuff and were probably beneath 
the floor plaster. This particular group contains 20-25 pits. 



included, which reduces the number of cavate 
wall depressions considered from 46 to 26. M- 
60 is the "type room" for this complex: it has all 
the features except a slot. In M-60 there would 
be two wall depressions to go with the two 
metate rests except that the "depression" is so 
large as to have been considered a floor-level 
niche. 

The low frequency of these features in 
combination with the incomplete knowledge of 
floors (which especially affects counts of ridges 
and metate rests) makes statements about co- 
occurrence of these features tentative. Still, the 
small sample suggests that slots and floor ridges 
are likely to co-occur: five of seven slots in 
cavates were in rooms that also contained ridges. 
Wall depressions near the floor are much more 
common than metate rests, but four of six 



metate rests were associated with wall 
depressions. The associations between ridges 
and slots, and between metate rests and wall 
depressions are stronger than those outside these 
pairs; we encountered no chambers containing 
all four feature types (Table 4.20). Nonetheless, 
ridges, slots, and metate rests are more likely to 
be found in chambers having one of the other 
features types than they are to be found alone. 
If ridges and slots were in fact somehow related 
to storage, this finding suggests that storage and 
milling may have often taken place in different 
rooms. 

Loom Anchor (Code 13) 

In a surprising number of cases, we 
found wooden loops remaining in floors (Figures 
4.8, 4.9). When enough floor is visible, these 



FEATURES 145 






Figure 4.6. Complex of features in M-40 including the best example of a metate rest we observed. 
Note the shallow plaster basin at the base of the incline and the floor ridge passing 
across the room directly in front of the metate rest. There is a groove in the wall where 
the floor ridge intersects the wall which is suggestive of a "slot. " 



features are usually found in groups of four or 
more; Hewett (1909b) says six to seven, and 
Peckham (1979:68) says six or more (see 
Figures 3.1, 4.8). Some form of cementing 
material is sometimes visible around the wood 
remnants. A. V. Kidder thought that the 
material used at Pecos was tamped-down ash 
(see Smith 1972:123); Peckham suggests 
compacted ash or clay. We thought it might be 
some form of clay, but we did not analyze the 
material. 

Several attributes support the contention 
that these features are loom anchors. As noted, 
they usually occur in straight lines. In several 
cases they are found in chambers with distinctive 
beam holes and other ceiling features. They are 



also very similar to features interpreted in this 
way in the western Anasazi/Western Pueblo 
area, where there is some ethnographic 
continuity in their use (see Smith's summary and 
description, 1972:121-123; Kent 1983a: 119- 
125). Like the examples described by Smith, 
several of these features are parallel to a 
chamber wall, which explains why they still 
exist in rooms that are heavily visited (several 
rooms at Tsankawi have floor remaining around 
the edge of the chamber with a worn area in the 
center). The most visible example we recorded, 
however, violates this pattern and runs more or 
less diagonally across the room (Figure 4.8). 

Though we observed several instances of 
loom anchors, we found few whole sets (M-59 



146 CAVATE STRUCTURES 





Figure 4.7. Grinding complex in room M-60. a. Note the side-by-side plaster 
basins similar to that seen in M-40 (Figure 4. 6). An inclined plaster 
line leading into the basin by the wall can be seen on the wall. The 
wall depression on the wall on the right is worn through the plaster, 
and a smaller, shallower plaster defect can be seen next to the 
deeper depression, b. B. Mills demonstrating how a grinder's feet 
could have formed a depression in the wall ("mano "=30 cm). This 
room also contained rock art (Figure 4.28) and is separated from M- 
59 (Figures 4.8, 4.15) by a combination tuff and masonry wall. 



Table 4.19. Metate Rest Occurrence by Location and Group. 



FEATURES 147 



Shape 



Group A Group F Group I Group M Tsankawi 



Total 



Floor 











4 





4 


Right wall 





1 











1 


Back wall 








1 








1 


Total 





1 


1 


4 





6 



Table 4.20. Co-occurrence of possible mealing complex features. 



Feature Types per 
Room 


Ridge 


Slot 


Wall 
Depression 


Metate 
Rest 


Sum of 
Features 


Occurring in 
Cavates (n) 


Single type 


5 


1 


20 


1 


27 


22 


Two feature types 














Ridge and slot 


3 


4 






7 


3 


Ridge and metate rest 


1 






1 


2 


1 


Wall depression and 
metate rest 






2 


1 


3 


1 


Three feature types 














Ridge, slot, 
wall depression 


1 


1 


1 




3 


1 


Ridge, metate rest, 
wall depression 


1 




1 


2 


4 


1 


Metate rest, slot, 
wall depression 




1 


2 


1 


4 


1 


Totals 


11 


7 


26 


6 


50 


30 



Note: Cavates only; only wall depressions 15 cm or less above floor. Columns show number of 
occurrences of a feature and the number of cavates in which the features occur. For example, three 
ridges co-occur with four slots (seven features) in three rooms. 



148 



CAVATE STRUCTURES 




Figure 4.8. Row of six loom anchors in cavate 
M-59; portions of the wooden loops are present in 
several of the holes and the use of different clay 
material to cement the anchors is visible in the 
hole in the center foreground. This room has a 
small back chamber, large floor-level niche, floor 
depressions, a slot (Figure 4.15) and other wall 
features, but no rock art. It was designated as a 
habitation room. 



Figure 4.9. Close-up of a wooden loom anchor loop 
in TS-59. This is a very large room with a full 
complement of "fdva" features and elaborate rock 
art. This loop is in a remarkable state of 
preservation in spite of heavy visitation to the 
room, because it (and other loom anchors 
observed) is located just at the edge of the heavy 
traffic area of the floor. It was covered when we 
worked here but was exposed and damaged by 
vandals sometime between August, 1986, and 
April, 1987. 




FEATURES 



149 



and TS-65 contain the most nearly complete 
recorded sets). From the cases we observed, the 
looms appear to have been more than 1 m wide, 
but how much more is not clear. Hewett's maps 
(1909b:660-665) show 1.35-2.10 m for five sets 
of six to seven anchors, averaging 1.6 m 
(s.d.=0.3). Six aligned probable anchors in TS- 
65 form a line 2.2 m long, but they may not all 
be contemporaneous, and subsets of four or five 
suggest a width of 1.1-1.5 m. Peckham 
(1979:67) illustrates some kivas (not cavate 
"kivas") containing two to four sets of these 
features. He notes several arrangements, 
including location by the wall similar to those in 
the cavates: 

Although often occurring as single 
alignments, pairs of loom hole 
alignments may be on either side of the 
hearth, paralleling the east-west axis of 
the kiva, or at more acute angles to this 
axis, or close to and roughly parallel to 
the kiva wall on either side of the 
ventilator opening. Multiple sets of 
loom holes frequently occur in the same 
general location indicating that they 
occasionally had to be replaced. 
(Peckham 1979:68) 

The traditional Pueblo looms shown by 
Kate Peck Kent are apparently held down by 
rocks or heavy logs, though one Hopi example 
may employ some form of anchor in the ground 
(1983b: 11, 33, 59, 78). Kent also provides a 
drawing of a prehistoric loom employing loom 
anchors (1983a: 120). Several of the looms she 
shows are located outdoors. 

All the loom anchors are presumably 
some form of cylinder in section, but for most 
only a plane shape is observable (Table 4.21). 
Their absence at Groups A and F is probably 
sampling error, though upper loom supports are 
also absent in our samples from those groups. 
Where there is one loom anchor, as at Group I, 
there are surely others. The one there was 
suggested by upper supports and the wall 



features in the room, and only a small area of 
the shallow fill was cleared. 

The diameter of loom anchors seems 
quite consistent. The unfortunate recent 
vandalism of an anchor at Tsankawi shows that 
they may be quite deep (at least 15 cm) and very 
regular cylinders below floor level. The 
construction apparently entailed boring a hole in 
the tuff floor of the chamber, inserting a bent 
twig (1-2.5 cm in diameter), and cementing the 
twig in place using a clayey substance. In the 
few cases observed, the clay is an olive-yellow 
color distinct from the surrounding tuff. 

Step (Code 35) 

In a few cases room floors appeared to 
have more than one level, and the feature 
dividing the levels was called a step. Two 
examples were recorded at Group M and one at 
Group I. It is quite conceivable that some no 
longer discernible room partition may have 
existed next to some steps. Two steps were also 
recorded in Group F (F-47); rather than dividing 
a room, however, these steps appear to have 
been stairs cut in the sloping back wall of a 
room in order to reach the higher cavate room 
(F-l) behind it. 

Axe Groove (Code 50) 

Seven examples of this feature were 
recorded at Tsankawi. The softness of the tuff 
raises some doubt that it would be effective in 
sharpening an axe made of hard stone, but an 
abrasive lap would be generated. Given erosion 
and location this functional assignment is 
speculative. 

Adobe Collar (Code 52) 

A single example of this feature "type" 
was recorded at Group F (F-37). It consists of 
an unburned semicircle of adobe that abuts the 
back wall just opposite the exterior door of a 
small chamber (Figure 4.10). The wall plaster 



150 



CAVATE STRUCTURES 



Table 4.21. Loom Anchor Occurrence and Dimensions. 

A. Loom Anchor Occurrence by Shape and Group 



Shape 



Group A Group F Group I Group M Tsankawi 



Total 



Oval 











1 


1 


2 


Round 








1 


4 


15 


20 


Cylindrical 











7 





7 


Total 








1 


12 


16 


29 



B. Mean Dimensions by Grouped Shape in Meters 



Dimension 


n 


Mean 


Standard 
Deviation 


Minimum 


Maximum 


CV 


Diameter 
Depth 


24 
5 


0.09 
0.03 


0.019 
0.029 


0.05 
0.01 


0.12 
0.08 


22.3 
89.5 



above the feature is worn and is also unburned. 
The collar is 4 cm high, and the diameter of the 
feature is 64 cm at the wall. Conjecturally, the 
enclosed area might have been used to hold pots. 

Deflector (Code 44) 

Deflectors in cavates are low walls next 
to the exterior opening or door of the chamber. 
The firepits we observed are nearly all in the 
same location, but not every firepit has a 
deflector associated with it. We recorded only 
three deflectors, one at Group I and two at 
Tsankawi, though the best example observed is 
in the upper, unrecorded part of Group F 
(Figure 4.11). 

Wall Features 

Large Floor-level Niche (Code 8) 

Large niches either at floor level or 
slightly below it seem to be relatively common 



in Bandelier cavates (Figure 4. 12, Table 4.22). 
The high counts are partly because the niches 
are large enough to be recognizable even in 
heavily eroded walls. Still, the niches were 
clearly part of the majority of cavates likely to 
have been used for habitation. They vary 
substantially in depth; the shallowest ones are 
borderline wall depressions, and the largest 
might be considered small back chambers (note 
that the maximum volume is greater than the 
minimum volume given for chambers). The 
basis for calling such a feature a niche rather 
than a chamber was primarily morphological. 
Niches are not constricted at the opening by a 
definable door; still, in some cases it was a 
matter of judgment whether to call them 
chambers or niches. There is little concrete 
evidence concerning their function; a few 
contain other wall features, but most do not. 
They are frequently smoked in the top, and often 
plastered, though the plaster coats may differ 



FEATURES 151 




Figure 4.10. The sole "adobe collar" observed by our crew. The collar is opposite the door in the 
small chamber of F-37. The wall plaster is worn behind the collar; a wall niche is 
present above to the left (scale =30 cm). 



from those on the wall in which the niche is 
located. 

For some reason, large floor-level niches 
are less often found on the left wall; we 
recorded approximately equal numbers of left 
and right walls, but nearly three times as many 
large niches are found on the right than on the 
left wall. They occur quite frequently in the 
back corners of the chamber, but corner 
locations do not show a similar preference for 
side (10 back right, 8 back left). The 
percentage of these features on back walls is 
somewhat higher than the percentage of back 
walls recorded might suggest, but since back 
walls are usually intact, the disproportion is not 
difficult to understand. This is reflected in the 



fact that one-fourth (33) of these niches recorded 
were located on noncavate back walls; clearly, 
in a room with three masonry walls and one tuff 
wall, the back, tuff wall would be the prime 
location for these handy features. The mean 
distance from the floor to the bottom of the 
opening of these niches is -3 cm. As would be 
expected from the feature name, about half were 
at floor level. 

These features tend to have rectangular 
openings, since a section of a cylinder is 
rectangular. The difference in shape assignment 
in the two most abundant categories is a function 
of the morphology of the backs of the niches: 
some are rounded, others more rectilinear. 



152 CAVATE STRUCTURES 




Figure 4.11. Well-preserved example of a deflector in upper Group F, Room 15 (this cavate was not 
recorded by the project). There is a basin firepit next to the chamber side of the 
deflector. 



The coefficient of variation for volume 
for this feature is very high, no doubt reflecting 
variability in function within features in this 
category. The largest case (an oval feature 
nearly 1.8 m tall at Group A) is almost twice the 
volume of the next largest example. The vast 
majority of large floor-level niches, however, 
fall at the smaller end of the scale, ranging from 
0.04 to 0.20 m 3 (Figure 4.13); features at 
Tsankawi and Frijoles follow similar 
distributions. Smaller examples could have 
accommodated a jar, for example, while some of 
the larger ones could have been for storing more 
items or food. Even when the extreme case is 
removed, the standard deviation for volume is 
greater than the mean (CV=112). In spite of 
the overall variability in volume, the mean 



volumes of this feature are remarkably similar 
across shapes, suggesting that although there 
may be several volume categories within the 
feature type, differing shape may not be a good 
indication of differing function. Over 80 percent 
of the relatively complete large floor level niches 
recorded are 0.2 m 3 or less in volume, but even 
within that group volume variability is high 
(CV=77.4), again suggesting a variety of 
functions for this feature type. 

Wall Niche (Code 9) 

Although this feature type might seem 
redundant with the large floor-level niche, the 
two seem to be distinct. Wall niches are 
generally much smaller than large floor-level 



FEATURES 153 





Figure 4.12. Large floor level niches. Lower photograph also shows wall 
replastering to specified height providing a distinct "dado'', a. An 
example of pairing of a small wall niche with a large floor-level 
niche in M-44. b. Example in A-47, with second large floor-level 
niche at the left edge of the photograph. Note the wear around the 
base of the opening. 



154 CAVATE STRUCTURES 



Table 4.22. Large Floor-level Niche Occurrence and Dimensions. 

A. Large Floor-level Niche Occurrence by Shape and Group 



Shape 



Group A Group F Group I Group M Tsankawi 



Total 



Rectangular solid 

Oval 

Trapezoidal 

Cylindrical 

Hemispherical 

Truncated cone 

Truncated pyramid 

Conical 

Spherical 

Irregular 

Total 



9 
3 


6 
1 
3 

1 


23 



12 





4 


1 





1 


1 


1 


1 


1 








7 


2 


11 





1 


1 


2 


3 


5 


1 











1 











1 



25 



24 



13 

4 
1 
1 
23 

2 

3 


47 



38 
9 

4 
2 

49 
3 

15 
1 
5 
1 

127 



Noncavate, grouped 







B. Chamber Location by Shape 



Shape 


Right 
Wall 


Back 
Wall 


Left 
Wall 


Exterior 
Wall 


Back 
Corners 


Front 
Corners 


Total 


Rectangular solid 


7 


22 


3 





5 


1 


38 


Oval 


2 


7 














9 


Trapezoidal 





3 


1 











4 


Cylindrical 


1 











1 





2 


Hemispherical 


7 


26 


4 


2 


6 


4 


49 


Truncated cone 





2 








1 





3 


Truncated pyramid 


2 


9 








4 





15 


Spherical 


3 


2 














5 


Irregular 














1 





1 


Total 


22 


71 


8 


2 


18 


5 


126' 



FEATURES 



155 



Table 4.22. (continued) 



C. Mean Dimensions by Shape in Meters 









Standard 








Shape 


n 


Mean 


Deviation 


Minimum 


Maximum 


CV 


All shapes 














Widths 


121 


0.57 


0.241 


0.12 


1.40 


42.4 


Heights 


116 


0.53 


0.280 


0.15 


1.78 


52.8 


Depths 


43 


0.48 


0.237 


0.11 


1.11 


49.6 


Volume 


110 


0.1281 


0.1707 


0.0039 


1.2582 


133.2 


Round 














Diameters 


56 


0.58 


0.261 


0.12 


1.40 


44.8 


Depths 


51 


0.50 


0.238 


0.15 


1.07 


44.7 


Volumes 


55 


0.1256 


0.1572 


0.0042 


0.6593 


125.1 


Rectangular 














Width 


33 


0.51 


0.203 


0.20 


1.04 


40.2 


Height 


33 


0.47 


0.175 


0.15 


0.58 


39.2 


Volume 


32 


0.0915 


0.0845 


0.0039 


0.3103 


92.3 


Pyramidal 














Base width 


14 


0.63 


0.215 


0.35 


1.10 


34.2 


Height 


14 


0.70 


0.201 


0.35 


1.11 


28.8 


Volume 


14 


0.1737 


0.1208 


0.0220 


0.4608 


69.6 


All height 


86 


-0.03 


0.107 


-0.30 


0.22 




above floor 















Note: Only niches considered to be more than 70 percent complete are included. C, round includes 
cylindrical, conical, and hemispherical; rectangular includes rectangle, bowed rectangle, and rectangular 
solid. 



"Conical niche is within another feature, which is in the right wall (not tabulated). 



156 CAVATE STRUCTURES 




FEATURES 157 



niches, they are usually some distance from the 
floor, and they tend to be shaped differently 
from the floor-level niches (examples may be 
seen in Figures 4.4, 4. 10, 4. 12a). Wall niches 
are often quite carefully shaped, through built-up 
plaster work around the opening, careful 
excavation into the tuff, or some combination of 
the two. 

The frequency and occurrence by group 
of wall niches are quite similar to those of the 
large floor-level niches (Table 4.23). Walls 
with large floor-level niches are probably more 
likely than others to have wall niches as well. 
The disproportionately low frequency of large 
floor-level niches on left walls is not matched by 
the wall niches, so that there is no one-for-one 
co-occurrence of the two niche types. We 
recorded 20 wall niches at or slightly below 
floor level, but the majority are 20-45 cm above 
the floor. Most are much smaller in volume 
than floor-level niches, with only a few reaching 
the mean volume for the latter (Figures 4.13, 
4.14). Thus the two niche types seem fairly 
distinct in both wall placement and size (based in 
large part on definition, of course), though they 
overlap in both respects. 

In some cases the two types of niches 
seem to be paired (Figure 4.12a). Given the 
similarity in counts, it is possible that these two 
niche types may have served complementary 
functions. Their co-occurrence in chambers and 
noncavates is shown in Table 4.24. Among 
rooms containing niches, niche occurrence can 
be divided approximately into thirds: chambers 
with wall niches only (37 percent), those with 
large floor-level niches only (35 percent), and 
those with both types occurring together (28 
percent). The Tsankawi cavates are distinctive 
in having some chambers with larger numbers of 
niches of both types; while two of each type is 
the maximum for our Frijoles sample, one 
Tsankawi chamber has three of each, and several 
others have multiple wall niches and a single 
large floor-level niche. 



Slot (Code 10) 

Slots are somewhat enigmatic features. 
They are usually incised vertical ovals with their 
bases near the floor (Figure 4. 15). Where the 
floor is visible, floor ridges seem to be 
associated with slots, but the number of cases is 
small (Table 4.20). These features seem to have 
been intended to hold the end of a plank, 
thereby forming low partitions of the room, 
perhaps for storage bins. 

Although there are only 10 slots in the 
sample, their distribution has noteworthy aspects 
(Table 4.25). Slots are absent at Tsankawi, and 
half of those recorded are at Group M. There 
also seem to be more than might be expected in 
the right wall of the chamber. Three of the 
recorded examples are on noncavate back walls, 
reducing the opportunity to observe co- 
occurrence with floor features. Though 
variable, the dimensions could all accommodate 
planks or perhaps a slab (see Carlson and Kohler 
1989:55). With the exception of the highest 
value of 25 cm, the mean height above the floor 
of observed slots is close enough to the floor to 
have held a separator for a substance such as 
grain, if the floor had been built up (as with a 
floor ridge). 

Viga Hole (Code 11) 

The assignment of several types of wall 
holes to feature categories involved some degree 
of subjectivity. Features were called viga holes 
if they were of sufficient size to contain a 
substantial beam and were located in such a way 
that roof support was likely. Particularly on 
open back walls to former masonry rooms, rows 
of probable viga holes with little vertical 
separation suggest remodeling episodes. Size, 
shape, and angle were the primary criteria for 
this feature assignment; in all we observed over 
500 features we called viga holes. 

Other chamber locations include ceiling 
(five holes) and an overhang above the cavate 
(one set of six). Forty percent (1 14) of the viga 



158 CAVATE STRUCTURES 



Table 4.23. Wall Niche Occurrence and Dimensions. 

A. Wall Niche Occurrence by Type and Group 



Shape 



Group A Group F Group I Group M Tsankawi 



Total 



Rectangular 

Oval 

Cylindrical 

Hemispherical 

Truncated cone 

Truncated pyramid 

Conical 

Spherical 

Irregular 



8 
1 
6 
13 

1 






13 


3 


7 


3 





3 


4 





3 


7 


5 


10 








1 


1 


1 


1 








1 





















20 
8 
7 

24 



51 
15 
20 
59 

2 
5 
2 

1 
1 



Total 



29 



28 



26 



64 



156 



Noncavate, grouped 











B. Chamber Location by Shape 








Shape 


Right 
Wall 


Back 

wall 


Left 
Wall 


Exterior 
Wall 


Back 
Corners 


Front 
Corners 


Total 


Rectangular solid 


8 


34 


5 


2 


2 





51 


Oval 


2 


8 


4 


1 








15 


Cylindrical 


1 


14 


2 


2 


1 





20 


Hemispherical 


6 


34 


5 


6 


3 


1 


55 


Truncated cone 








1 


1 








2 


Truncated pyramid 





2 


3 











5 


Conical 





2 














2 


Spherical 








1 











1 


Irregular 





1 














1 


Total 


17 


95 


21 


12 


6 


1 


152 



Table 4.23. (continued) 



FEATURES 159 



Shape 



C. Mean Dimensions in Meters 

Standard 
Mean Deviation Minimum Maximum 



CV 



All volume 


141 


Round 




Diameter 


81 


Depth 


79 


Volume 


79 


Rectangular 




Width 


46 


Height 


46 


Depth 


29 


Volume 


46 


All height 
above floor 


115 



0.0248 

0.25 
0.24 
0.0219 

0.27 

0.25 

0.13 

0.0284 

0.31 



0.0567 



0.0001 



0.3881 



228.4 



0.181 


0.06 


0.90 


72.4 


0.221 


0.06 


1.30 


92.6 


0.0526 


0.0001 


0.3793 


240.2 


0.202 


0.06 


0.84 


73.8 


0.196 


0.04 


0.75 


78.2 


0.101 


0.05 


0.38 


77.2 


0.0457 


0.0002 


0.1718 


160.9 


0.260 


-0.21 


1.48 


83.9 



160 CAVATE STRUCTURES 



< 





,** 




tn <o 




©.* 




o£ 




: E8 




ILH 


0) 


• <3 







o 




z 




75 




£ 





CO 

d 

CM 

1- 

ci 



s 

d 

CO 

o 



d E 

(0 3 
§ P 






< 



• • 



•^^^y*^^ 



-8 



o 
d 

CO 

o 



o 



o 
d 



i ■ i ■ i ■ i ■ i ■ i ■ i ■ i ■ i ' i ■ i ■ i 

T^qa>cqN(qu>«tcocMT^q 
r^-^dddddddddd 



too\i 3Aoqe m6;8H 



S 

d 



o 
■ 



*: 






Si 



FEATURES 



161 



Table 4.24. Co-occurrence of Floor-level and Wall Niches by Group. 



Rooms with: 



Group A Group F Group I Group M Tsankawi Total 



Floor-level niche only 


10 


11 


6 


8 


19 


54 


Wall niche only 


13 


16 


6 


10 


13 


58 


One of each type 


2 


2 


2 


6 


2 


14 


Multiple of each 


2 


1 





1 


1 


5 


Floor-level > wall niches 


2 


3 





2 


3 


10 


Wall > floor level niches 


2 


1 





2 


9 


14 



Total rooms with niches 



31 



34 



14 



29 



47 



155 



Total floor-level niches 


23 


25 


8 


24 


48 


128 


Total wall niches 


29 


28 


9 


26 


64 


156 


Maximum n floor-level 


2 


3 


1 


2 


5 




niches 














Maximum n wall niches 


2 


2 


2 


2 


7 





holes we recorded individually are associated 
with noncavate back walls, and almost as many 
viga holes were recorded in sets as individually. 
The combined counts demonstrate the 
importance of cliff-backed masonry rooms at 
these sites, particularly Groups A and F. The 
relatively high frequency of viga holes at Group 
A fits with the large number of "back wall" 
features there (Table 4.26). 

Why use vigas in a cavate? Viga holes 
are present in "cavates" largely because many 
cavates had masonry closing structures. Also, 
the counts by location show only individually 
recorded viga holes, not those recorded in 
groups on noncavate forms. Still, features that 
are to all appearance viga holes do occur in 
closed cavates, suggesting either some ceiling or 
rack structure. Some of these may have been 
upper loom supports lacking other recognition 



criteria (Figure 4.24). Many such features have 
smoke-blackened interiors, suggesting that they 
may have served only to hold some form of 
beam for part of the period of use of the cavate. 

Figure 4. 16 gives an indication of the 
variation in viga hole height, diameter, and 
depth. There is a concentration of holes 10-15 
cm in diameter, 15-25 cm in depth, and 1.2-1.6 
m high. We recorded height above floor only 
for cavates; if we had included noncavate back 
walls, the mean would be larger. The height 
above floor is fairly consistent; indeed, some 
height above the floor is necessary for a hole to 
be assigned to this feature type. Although a few 
cases fall in the 1.8-2.0 m range that most 
moderns would consider a comfortable ceiling 
height, 56 percent of the included cases fall 
between 1.2 and 1.6 m above the floor (Figure 
4.16). Given the values below 1 m and the 



162 CAVATE STRUCTURES 




Figure 4.15. Good example of a slot in cavate M-59 (scale =30 cm). This room also contains loom 
anchors and a large floor level niche, but no recognizable milling features. 



placement inside cavates, it is likely that if some 
of these viga-shaped holes did in fact contain 
vigas, the beams had functions other than ceiling 
support. 

Possible Latilla Hole (Code 12) 

This feature type is more ambiguous 
than the viga hole. Size and placement again 
contribute to its use. Occasionally series of 
small holes are found at about ceiling height, 
which may have been for latillas; these are the 
ideal case, but the category also includes less 
clear examples (Table 4.27). 

The feature category was originally 
inspired by a structure by the Tsankawi trail, 



where there is a series of 10 closely spaced, 
flattened, hemispherical (or "ovals with depth") 
holes that clearly supported a roof above an area 
in front of a cavate (probably Lister's C-119). 
After all the present measurements had been 
taken, I returned to this set of "ideal" latilla 
holes and measured them; they are 0.07-0.13 m 
wide, 0.064.11 m high, and 0.06-0.09 m deep. 
They are thus somewhat larger than the average 
for other features recorded in this category, 
though the depths are similar. Figure 4.17 
shows that the features placed in this category 
are generally smaller and shallower and the 
measurements more dispersed than viga holes 
(compare Figure 4.16). Given the mean viga 
hole diameter and that for possible latillas, it 
appears that the wood being used in cavates 
tended to be fairly small. Comparing the height 



FEATURES 



163 



Table 4.25. Slot Occurrence and Dimensions. 

A. Occurrence by Type and Group 



Shape 


Group A 


Group F 


Group I 


Group 


M 


Tsankawi 


Total 


Oval 


1 
















1 


Linear 


2 





1 










3 


Rectangular 





1 





5 







6 


Total 


3 


1 


1 


5 







10 



Shape 



B. Chamber Location by Shape 

Right Wall Back Wall 



Left wall 



Total 



Rectangular 

Oval 

Linear 



4 


1 


1 


6 





1 





1 





3 





3 



Total 



10 



C. Mean Dimensions of Oblong Shapes in Meters 









Standard 








Dimensions 


n 


Mean 


Deviation 


Minimum 


Maximum 


CV 


Width 


10 


0.06 


0.027 


0.02 


0.11 


42.4 


Height 


10 


0.26 


0.080 


0.14 


0.40 


30.7 


Depth 


3 


0.04 




0.02 


0.07 




Height 


5 


0.14 


0.14 


-0.07 


0.25 


98.4 


above floor 















Note: C, oblong includes rectangular, oval, trapezoidal, and triangular shapes. Only features judged 
to be more than 70 percent present are included. All available cases and shapes are included in 
height-above-floor measurements. 



164 CAVATE STRUCTURES 

Table 4.26. Viga Hole Occurrence and Dimensions. 

A. Viga Hole Occurrence by Shape and Group 



Shape 



Group A Group F Group I Group M Tsankawi 



Total 



Rectangular 

Oval 

Cylindrical 

Hemispherical 

Truncated cone 

Conical 



3 
1 
59 
34 





2 








1 








11 


8 


42 


17 


8 


7 


4 


1 


5 


2 


2 






2 
1 

19 

34 

9 

9 



7 

3 

139 

100 

19 

13 



Total 



97 



37 



19 



54 



74 



281 



Noncavate, grouped 



67 



87 



48 



26 



234 



B. Chamber Location by Shape 





Right 


Back 


Left 


Exterior Wall 


Back 






Shape 


Wall 


Wall 


Wall 


& Corners 


Corners 


Other 


Total 


Rectangular 


1 


4 


2 











7 


Oval 





1 


1 








1 


3 


Cylindrical 


15 


91 


15 


3 


7 


8 


139 


Hemispherical 


7 


77 


5 


4 


5 


2 


100 


Truncated 


2 


9 


2 


2 


4 





19 


cone 

















Conical 



13 



Total 



26 



186 



27 



10 



21 



11 



281 



/ 



FEATURES 



165 



Table 4.26. (continued) 



Dimension 



C. Mean Dimensions by Grouped Shape in Meters 



Mean 



Standard 
Deviation 



Minimum Maximum 



CV 



Diameter 


255 


0.14 


0.056 


0.04 


0.38 


40.5 


Depth 


251 


0.16 


0.097 


0.03 


0.71 


62.2 


Volume 


245 


0.0026 


0.0038 


0.00003 


0.0431 


148.5 


Height 


121 


1.39 


0.266 


0.85 


1.94 


19.2 


above floor 















Note: Shapes included are cylindrical, hemispherical, conical, and truncated cone. Only features 
judged to be more than 70 percent complete are included. Heights are given only for chambers with 
15 cm or less of fill. 



measurements with those for viga holes shows 
that this feature class is on average lower than 
viga holes, contrary to what would be expected 
if "latillas" were placed on top of vigas. In only 
a few cases did these two types of holes occur 
together in the "proper" sequence; of 150 
cavates or noncavates where either feature type 
occurs, only 10 have both types. As with viga 
holes, we can only speculate about the actual 
function of apparently structural holes inside 
cavates. Holes for small beams do seem to have 
been used, and it seems more likely that they 
supported shelves than the prehistoric equivalent 
to false ceilings. 

Beam Seat (Code 30) 

This code differs from the viga hole 
primarily in that placement and pairing do not 
suggest roof support. Size still suggests that a 
substantial piece of wood would have been 
inserted (Figure 4.18a). The distinction is 
somewhat subjective, and some crossover 
between the two categories would be likely if we 
were to reclassify all cases. 



Other than a low frequency at Group F, 
there is little remarkable in the distributions of 
beam seats (Table 4.28). Compared to viga 
holes, features recorded as beam seats are much 
more often oval or rectangular, though the 
majority of openings are round; the mean 
diameter for beam seats is about two-thirds that 
for viga holes, and they are somewhat shallower 
on average (Figure 4.18b). Beam seats are 
more evenly distributed on chamber walls than 
are viga holes, though the majority are again 
found on back walls. Although a few possible 
beam seats were located near the chamber floor, 
around 70 percent are 0.8-1.4 m above the 
floor, with most of those in the 0.8-1.0 m range. 
Figure 4.18a shows a concentration of beam 
seats around 9 cm in diameter, 20 cm in depth, 
and 60-100 cm above the floor. 

Indeterminate Hole (Code 28) 

Cavates and noncavates have a great 
many small holes in their walls, and since 
almost none of the holes now contains 
identifiable remains, it is impossible to know 
what their function was, resulting in heavy use 



166 CAVATE STRUCTURES 



<D 


z 

CO 




9'V O'V 90 















FEATURES 



167 



Table 4.27. "LatUla " Hole Occurrence and Dimensions. 

A. "LatUla " Hole Occurrence by Shape and Group 



Shape 


Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


Round 











3 





3 


Cylindrical 


10 


2 





3 


8 


23 


Hemispherical 














1 


1 


Conical 


11 


3 








7 


21 



Total 



21 



16 



48 



Noncavate, grouped 



10 



15 



B. Chamber Location by Shape 





Right 


Back 


Left 


Exterior Wall 


Back 






Shape 


Wall 


Wall 


Wall 


& Corners 


Corners 


Other 


Total 


Round 





3 














3 


Cylindrical 


2 


14 


4 


3 








23 


Hemispherical 





1 














1 


Conical 


6 


8 


2 


3 


1 


1 


21 


Total 


8 


26 


6 


6 


1 


1 


48 



C. Mean Dimensions in Meters 



Dimension 


n 


Mean 


Standard 
Deviation 


Minimum 


Maximum 


cv 


Diameter 


46 


0.04 


0.018 


0.02 


0.08 


39.7 


Depth 


43 


0.09 


0.037 


0.03 


0.20 


41.4 


Volume 


45 


0.00013 


0.00016 


0.00001 


0.0007 


128.9 


Height above 
floor 


36 


1.09 


0.303 


0.40 


1.64 


27.8 



Note: C, only cases more than 70 percent complete are included. Height measurements are only from 
chambers with 15 cm or less of fill. 



168 CAVATE STRUCTURES 



0) 


O 

X 




S'V £V SO 90 



■s •-' ? *» 

K « £ S{ 
^ 5 s ^ 

III i 

I « s ? 



^> 






^J 



*> "3 



1 



Si 



FEATURES 



169 




170 



CAVATE STRUCTURES 



Table 4.28. Beam Seat Occurrence and Dimensions. 

A. Beam Seat Occurrence by Shape and Group 



Shape 


Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


Oval 


12 











3 


15 


Rectangular 


7 





2 


4 


19 


32 


Cylindrical 


6 





3 


16 


27 


52 


Hemispherical 


12 


1 


2 


11 


13 


39 


Truncated cone 


1 


3 


2 


3 


7 


16 


Truncated pyramid 











3 


1 


4 


Conical 


1 


3 





4 


4 


12 


Irregular 











3 


2 


5 



Total 



39 



44 



76 



175 



Noncavate, grouped 



B. Chamber Location by Shape 





Right 


Back 


Left 


Exterior Wall 


Back 






Shape 


Wall 


Wall 


Wall 


& Corners 


Corners 


Ceiling 


Total 


Rectangular 


12 


6 


7 


5 


1 


1 


32 


Oval 


3 


6 


3 


1 





2 


15 


Cylindrical 


10 


21 


13 


5 


2 





51 


Hemispherical 


5 


19 


3 


7 





2 


36 


Truncated 


2 


9 


4 





1 





16 


cone 
















Truncated 


1 


1 


1 











3 


pyramid 
















Conical 


3 


6 








1 


2 


12 


Irregular 


3 





2 











5 


Total 


39 


68 


33 


18 


5 


7 


170 



FEATURES 



171 



Table 4.28. (continued) 



C. Mean Dimensions by Shape in Meters 









Standard 








Shape 


n 


Mean 


Deviation 


Minimum 


Maximum 


CV 


Round 














Diameter 


114 


0.09 


0.035 


0.06 


0.90 


72.4 


Depth 


114 


0.14 


0.140 


0.01 


1.15 


101.6 


Volume 


113 


0.0010 


0.0023 


0.00001 


0.0208 


116.2 


Rectangular 














Width 


29 


0.13 


0.051 


0.04 


0.24 


38.4 


Height 


29 


0.09 


0.045 


0.06 


0.17 


30.8 


Volume 


29 


0.0010 


0.0009 


0.0001 


0.0040 


96.7 


Height 


137 


1.01 


0.034 


0.01 


1.78 


33.2 


above floor 8 















Note: B, two features of unknown shape and three located within other features are not shown. 
"Heights above floor from chambers with less than 15 cm fill only. 



of this feature type during recording. This 
category certainly includes features that had a 
wide variety of functions, as well as some holes 
that are either natural or postoccupational. 
While they often occur in groups at similar 
heights above the floor, they are distinguished 
from possible latilla holes in arrangement and in- 
ability to support a small beam; there is again 
overlap, as measurements and multivariate 
analyses clearly show. Most are fairly small, 
but a few large, truly indeterminate holes are 
also included in this category (Table 4.29). 

The numbers of walls and features 
recorded in the five study areas can be used as 
a means for estimating how many walls would 
be expected at a group given an even or random 
distribution of wall features on walls. At 
Tsankawi there are considerably more holes than 
"wall expectation" and fewer at Groups A, I, 



and M. Some of this difference may result from 
differences in tuff at Tsankawi as compared to 
Frijoles: the top stratum at Tsankawi contains a 
great number of vesicles. We tried to be 
conservative in what we recorded as features 
(indeed we had to be, given the number of holes 
in some walls), but the tuff may have inflated 
the hole count. The poorer preservation of 
plaster at Tsankawi may also add to the higher 
count there, since fewer natural and/or disused 
holes remain covered by plaster. 

The measurement data show that this 
feature category does indeed cover a variety of 
holes in the wall, including some very large 
ones (Table 4.29, Figures 4.19a, b-4.20a, b), 
and that they are located at all heights. The 
means, however, show that most of these holes 
are fairly small and tend to occur a little less 
than 1 m above the floor. Modally (Figures 



172 CAVATE STRUCTURES 




FEATURES 173 



Table 4.29. Indeterminate Hole Occurrence and Dimensions. 

A. Indeterminate Hole Occurrence by Type and Group 



Shape 



Group A Group F Group I Group M Tsankawi 



Total 



Rectangular 

Oval 

Circular portion 

Cylindrical 

Hemispherical 

Truncated cone 

Truncated pyramid 

Conical 

Irregular 



2 


1 


3 


1 


4 


5 


1 


2 


1 


5 





2 


43 


22 


29 


49 


26 


20 


13 


10 





1 


1 


3 


1 





2 


1 


14 


24 


2 


5 











1 



26 

9 

5 

147 

39 
8 


41 
2 



33 

21 

13 

290 

108 

13 

4 

86 

3 



Total 



91 



78 



51 



74 



277 



571 



Noncavate, grouped 



15 



10 



10 



37 







B. 


Chamber Location by Shape 








Shape 


Right 
Wall 


Back 
Wall 


Left 
Wall 


Exterior Wall 
& Corners 


Back 
Corners 


Ceiling 


Total 


Rectangular 


8 


7 


6 


11 








32 


Oval 


5 


10 


2 


3 





1 


21 


Circular 
portions 


2 


4 


2 


5 








13 


Cylindrical 


56 


119 


62 


35 


4 


11 


287 


Hemispherical 


15 


46 


33 


9 


2 





105 


Truncated 
cone 


5 


4 


3 








1 


13 


Truncated 
pyramid 





3 


1 











4 


Conical 


26 


29 


14 


9 


3 


5 


86 


Irregular 








1 





1 


1 


3 


Total 


117 


222 


124 


72 


10 


19 


564 



174 CAVATE STRUCTURES 



Table 4.29. (continued) 



C. Mean Dimensions by Grouped Shape in Meters 



Shape 


n 


Mean 


Standard 
Deviation 


Minimum 


Maximum 


CV 


Round 














Diameter 


489 


0.05 


0.037 


0.01 


0.40 


78.5 


Depth 


476 


0.08 


0.134 


0.01 


2.50 


170.1 


Volume 


472 


0.0008 


0.0144 


0.0000 


0.3142 


1812.5 


Rectangular 














Width 


33 


0.09 


0.056 


0.02 


0.32 


58.9 


Height 


33 


0.06 


0.024 


0.02 


0.12 


42.4 


Volume 


32 


0.0004 


0.0005 


0.00002 


0.0026 


130.0 


Oval 














Width 


21 


0.10 


0.523 


0.03 


0.24 


51.1 


Height 


21 


0.12 


0.151 


0.01 


0.60 


125.2 


Volume 


11 


0.0007 


0.0014 


0.00004 


0.0050 


188.5 


All height 
above floor 


411 


0.86 


0.399 


-0.05 


2.32 


46.2 



Note: B, seven cases in floor, unknown feature and unknown part are not included. C, includes only 
features that are 70 percent or more present. Round includes cylindrical, conical, and hemispherical 
shapes. Height above floor is only from chambers with less than 15 cm of fill. 



4. 19a, b), indeterminate holes are only 3 cm in 
diameter and 5 cm deep, and while the various 
sizes are fairly evenly split between Tsankawi 
and Frijoles, there is some tendency for smaller 
holes to be found at Frijoles and larger ones at 
Tsankawi (Figure 4.20a). Of 370 cases that met 
the criteria for inclusion in Figure 4.20a, 323 
(87%) had diameters and depths of 10 cm or 
less. Even with removal of cases with extreme 
diameter and depth values, the mean diameter 
remained at 4 cm, and the mean height above 
floor remained at 84 cm. The mean height 
above floor is very close to what might be 
considered a "typical" plaster height in the 
cavates, and many smaller indeterminate holes 
do occur at the plaster line. In that location it is 
easy to visualize them as containing pegs on 



which things (canteens, clothes, and so forth) 
could be hung, and once in a while they do 
contain sticks. Another indication that some 
may have been for pegs is that the plaster 
around the openings is often broken away, 
presumably from removal of the peg. 

Multivariate Analyses of Holes in Walls 

As an alternative means of examining 
the many round holes in cavate walls, two 
multi variate techniques were applied to a set of 
features conforming to the following 
specifications: 

Feature types: viga holes, latilla (?) 
holes, beam seats, and indeterminate holes 



FEATURES 175 



(0 
0) 

o 

X 

o 

5 

•c 

E 
o 

"O 

c 




I ' I ' I ■ I ■ I ' I ' I ■ I ' I ' I ' I ■ I 

8 00QOOOQQOOOO 
r- ^ t- 

tunoo 



> *a 6 ^ 

§ 13 $ .5 




S 






176 CAVATE STRUCTURES 




FEATURES 177 




178 CAVATE STRUCTURES 



CO 
0) 



c 

• MB 

E 

■o 
c 

"55 

73 

E 




*v z\ ao o 



1 I 

f 



I 



I 



* 



•v* ««, «*i 



8 



.60 



FEATURES 



179 



Shapes: round orifices, cylindrical, 
conical, truncated cone, and spherical (a 
maximum diameter of 40 cm was used to 
exclude a few very large cases) 

Other criteria: only features judged to be 
at least 70 percent complete and located in 
chambers having less than 25 cm of fill (to 
control for completeness and accurate height 
above floor) 

This pruning procedure and exclusion of features 
with missing values resulted in a group of 650 
holes. 

Cluster Analysis 

A cluster analysis (SAS FASTCLUS 
procedure) was run requesting six clusters, based 
on diameter, depth, and height above the floor. 
Repeated passes were made through the data 
readjusting the cluster "seeds" and locating cases 
in the closest cluster. The six clusters identified 
by this analysis vary from 5 to 242 members 
(Table 4.30). Six clusters were requested, to 
allow the program to identify two clusters more 
than the four categories presumably present. 
Although the clusters do not correspond 
precisely to the feature types, the fact that 
cluster membership is greatly reduced by cluster 
6 suggests that searching for more than six 
clusters would be inappropriate. Predictably, 
indeterminate holes are spread the most evenly 
over the most clusters, though two of the 
clusters comprise mostly features from this 
category. Viga holes also dominate a cluster, 
though they are split between two clusters. 

The groups of holes created by this 
analysis (it should be remembered that the 
feature types were not provided to the cluster 
analysis) are recognizable by their means and 
form what may be useful-even functional- 
subdivisions of circular holes. Thus cluster 1, 
the group with the most members, consists of 
holes 2-8 cm in diameter located near the top of 
the plaster in many rooms. Cluster 2 contains 



considerably larger and deeper holes much 
higher on the wall, and cluster 3 is also well up 
the wall, though the holes are smaller, 
shallower, and lower. Cluster 4 is composed of 
much smaller holes lower on the wall. Clusters 
5 and 6 are uncommon shapes and locations; 
both are deep, especially cluster 6, but cluster 6 
is quite high (similar in height to cluster 3) and 
cluster 5 is near the floor. As might be 
predicted, the holes identified as viga holes are 
fairly consistent, though they come in two sizes. 

Discriminant Analysis 

This large group of round hole features 
was also analyzed using discriminant analysis 
(the SAS DISCRIM procedure; 645 cases 
analyzed, 5 omitted due to missing values). For 
the discriminant analysis the program was 
provided with types assigned and then calculated 
a profile for each feature type. Once again the 
variables used to describe the features were 
diameter, depth, and height above floor. The 
individual cases were then compared to the 
profiles and placed in the one to which they 
most closely conform. For this analysis the 
prior probability that a feature would fall into a 
given type was set at equal, which is not the 
case for the actual distribution, since 
indeterminate holes form 62 percent of the total. 
Because it is of interest whether or not 
"indeterminate" holes form an identifiable 
category, this is a reasonable prior condition. 
Based on a test of covariance matrix 
homogeneity performed by the program, within 
covariance matrices were used in the 
discriminant function. 

The discriminant analysis gives an idea 
of the metric overlap among the feature types 
(Table 4.31, Figures 4.18b, 4.21). Thus, the 
majority of features called viga holes and beam 
seats are described by similar measurements, 
though there are probably at least two 
subgroups: one higher, larger, and deeper (viga 
hole means) and the other lower (about 1 m) and 
smaller. Each of these also overlaps with the 



180 CAVATE STRUCTURES 



Table 4.30. Results of Ouster Analysis on Round Wall Holes. 

A. Members and Means for Hole Clusters 

Cluster n Near Diameter Depth 



Height Above Floor 



1 


242 


3 


0.050 


0.081 


0.853 


2 


74 


3 


0.098 


0.180 


1.639 


3 


186 


2 


0.067 


0.099 


1.268 


4 


126 


5 


0.049 


0.062 


0.421 


5 


17 


4 


0.097 


0.196 


0.100 


6 


5 


3 


0.164 


0.838 


1.182 



B. Ouster Membership by Assigned Feature Type 



Cluster 


Viga 


Latilla 


Indeterminate 


Beam Seat 


Total 


1 


16 


18 


173 


35 


242 


2 


47 


1 


18 


8 


74 


3 


49 


15 


80 


42 


186 


4 





1 


117 


8 


126 


5 


1 





13 


3 


17 


6 


3 





1 


1 


5 


Total 


116 


35 


402 


97 


650 



Table 4.31. Discriminant Analysis Classification of Feature Types. 

Computer-Assigned Type 



Original Type 


Viga 


Latilla 


Indeterminate 


Beam Seat 


Total 


Viga hole 


80 


16 


1 


15 


112 


LatUla (?) 
hole 


1 


30 


4 





35 


Indeterminate 
hole 


12 


206 


158 


25 


401 


Beam seat 


27 


22 


12 


36 


97 


Total 


120 


274 


175 


76 


645 



FEATURES 



181 



0) 



a 

0) 

o 



0) 




ffV ?V 90 



S 5 



«u « <o 

§5 § 

111 
^ 3 ^ 

'E «■§ 

3*1 

111 

ill 



II 



5 .g e 

©> ?> *• 



£ .2 -^ 

1 J "1 

.2P - §o 

e 

5 
|l| 

° i * 
IS* 

©:-**■* 






182 CAVATE STRUCTURES 



other groups, which are smaller holes. Beam 
seat, as a category for features that are more 
difficult to interpret, shows considerably more 
overlap with the other two categories than does 
viga hole. The possible latilla hole group is the 
most consistent, probably at least in part because 
it is the smallest group. Indeterminate holes 
appear to fall into at least two main groups, 
more of which are like "latilla" holes than the 
smaller holes described by the measurements for 
the entire indeterminate hole group. 

Preservation and feature variability will 
always mean that a large number of wall 
features will have to be placed in a category 
such as "indeterminate holes," but as used here, 
this category is too inclusive. In future 
recording of features in cavates more types of 
small wall holes should be recorded, subdividing 
the "indeterminate hole" category used in this 
study. Additional types suggested by the 
analyses and by observation include: 

—peg holes. This type should be used 
for holes still containing portions of pegs or 
holes that show evidence for the removal of a 
peg. The group is likely to be fairly small, but 
isolating features with a known function will 
help place other, less easily interpreted features. 

-small holes at or near the top of the 
wall plaster. As noted, these may have been for 
pegs, but that use is less clear. It can be 
compared to the peg hole category. 

Possible Upper Loom Support (Code 14) 

The clearest examples of upper loom 
supports are deep, vigalike holes near the top of 
a chamber. The holes occur in pairs at a 
considerable angle to one another. Hewett's 
reconstruction of these features shows forked 
logs projecting from the seats; I do not know if 
this is based on intact examples or on 
speculation. Features associated with these 
angled supports include grooves in the ceiling 
(presumably to accommodate a cross-bar) and 



rows of loom anchors in the floor. As can be 
seen in Kent's (1983b) photographs of traditional 
looms, ordinary vigas can suffice as upper loom 
supports for looms close to walls, and it is quite 
possible that some de facto loom supports were 
called something else. The low ceilings and 
inclined walls of many cavates, however, may 
have required more specialized upper supports. 

The scarcity of identified upper loom 
supports at Groups F and A combined with the 
lack of loom anchors at those groups suggests 
that this pattern may be more than a sampling 
artifact (Table 4.32). Loom supports tend to be 
cylindrical holes in or near the ceiling of the 
chamber in which they are found. As we 
recorded these features, we came to recognize 
that in addition to their location, their angle of 
entrance into the ceiling was an important 
attribute. Because of the lateness of the 
realization and the difficulty of measuring it, 
however, we did not record this angle. The 
observed cases suggest that they are on average 
larger than viga holes and can be very deep. 

Smokehole (Code 15) 

Features called smoke holes are very 
commonly found near the chamber entrance. It 
seems likely that some smoke holes were 
separated from the door by a lintel, since some 
probable smoke holes now show as enlargements 
of the tops of doors, sometimes associated with 
grooves for rock or wood lintels. Smoke holes 
are usually quite large in diameter and angle up 
as they pass to the outside. Size, angle, and 
placement well up the wall are the most 
important criteria in assignment of this feature 
type. 

Smoke holes are slightly more abundant 
at Groups F and I than might be expected based 
on overall feature counts, and they are markedly 
infrequent at Group M (Table 4.33). This 
emphasizes the fact that Group M rooms made 
especially heavy use of masonry closing walls. 
The chamber locations show the strong (and in 



FEATURES 



183 



Table 4.32. Loom Support Occurrence and Dimensions. 

A. Loom Support Occurrence by Shape and Group 



Shape 


Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


Trapezoidal 











1 





1 


Rectangular 














1 


1 


Cylindrical 


1 





3 


2 


14 


20 


Hemispherical 














2 


2 


Truncated cone 














1 


1 


Conical 














3 


3 


Total 


1 





3 


3 


21 


28 



Shape 



B. Chamber Location by Shape 

Right Wall Back Wall Left Wall 



Ceiling 



Total 



Rectangular 





Trapezoidal 





Cylindrical 


3 


Hemispherical 





Truncated cone 





Conical 








1 

4 

1 






2 






1 



11 

2 

3 



1 
1 
20 
2 
1 
3 



Total 



17 



28 



C. Mean Dimensions of Circular Shapes in Meters 



Dimension 


n 


Mean 


Standard 
Deviation 


Minimum 


Maximum 


CV 


Diameter 
Depth 


26 
26 


0.13 
0.38 


0.070 
0.317 


0.06 
0.08 


0.35 
1.75 


48.2 
83.4 



184 CAVATE STRUCTURES 

Table 4.33. Smokehole Occurrence and Dimensions. 

A. Smokehole Occurrence by Type and Group 



Shape 



Group A Group F Group I Group M Tsankawi Total 



Oval 

Trapezoidal 

Triangular 

Rectangular 

Cylindrical 

Truncated cone 







11 
1 



3 


2 








1 


1 








7 


5 









2 
1 

1 
21 




7 

1 

2 

1 

46 

1 



Total 



12 



11 



25 



58 



Shape 


Exterior 
Wall 


Back 

Wall 


Left 
Wall 


Rectangular 
solid 











Oval 


7 








Triangular 


1 








Trapezoidal 


1 








Cylindrical 


35 


1 


4 


Truncated 


1 









B. Chamber Location by Shape 



Ceiling 



Left 

Exterior 

Corner 



Floor 



Total 




1 



4 




1 





1 









1 





1 

7 

2 

1 

46 

1 



cone 



Total 



45 



58 







C. 


Mean Dimensions for Circular Shapes in Meters 














Standard 








Dimension 


n 




Mean 


Deviation 


Minimum 


Maximum 


CV 


Diameter 


39 




0.23 


0.073 


0.06 


0.35 


32.2 


Depth 


39 




0.49 


0.226 


0.15 


1.01 


46.2 



FEATURES 



185 



a majority of cases necessary) preference for 
placement of smoke holes in the exterior wall. 
Clearly, they tend to be cylindrical holes bored 
through the cliff, but as with upper loom 
supports, the fact that most angle upward needs 
to be noted in addition to the recorded data. 

Vents (Code 16) 

Vents differ from smoke holes in two 
respects: they are horizontal rather than angled 
upward and they are generally found lower on 
the wall (see Figure 2.21b to left of door). Any 
hole passing through a wall that is not either a 
door or a smokehole was called a vent. 
Openings into other rooms as well as to the 
outside are included in this feature type. Like 
smoke holes, vents are predominantly cylindrical 
holes in the exterior wall. They, too, are 
infrequent at Group M and relatively abundant at 
Group F (Table 4.34). 

Groove (Code 31) 

Judging from differing size, location, 
and orientation, this feature category covers 
several probable functions. We observed some 
examples in ceilings, where they may have 
accommodated loom cross-bars; examples high 
on walls might have supported roofing ("wall 
ledge" would have been a more appropriate code 
for these); some may have helped support 
vertical partitions; and some may have been 
decorative. This code, then, describes a feature 
of questionable function (Table 4.35). 

Wall Depression (Code 37) 

Wall depressions are usually shallow 
concavities in tuff walls. Some of these features 
are plastered over, while others are clearly 
abraded through the plaster into the tuff. These 
features rarely have enough of a shelf at the base 
to hold anything (indeed, features with any shelf 
at all would probably have been called niches 
rather than wall depressions). Wall depressions 



are found at varying heights above the floor, but 
it seems likely mat those close to the floor in the 
vicinity of mealing features resulted from 
grinders' feet pushing against the wall (Figure 
4.7, Table 4.20). Mills made this observation 
in the field, independent of Chapman's 
reconstruction (Hewett 1909a:451) showing the 
same probable function. 

Wall depressions are somewhat more 
abundant at Groups A and F (including the 
noncavate observations) than would be expected 
given numbers of features recorded (Table 
4.36). These features also tend to be placed in 
back walls and corners and right walls as 
opposed to other chamber locations. The 
associated depression and metate rest in M-60 
are located in the back corner of the room 
(Figure 4.7). The great variability in size, and 
thus probably in source and function, is visible 
in the mean volumes and areas and their 
coefficients of variation (Table 4.36C). 

Wall Ledge (Code 41) 

Occasionally a substantial ledge is found 
near the top of a chamber. Most likely, these 
features were an alternative roof support to viga 
holes, though they are much less common (see 
Figure 2.16b above door). The examples we 
encountered did not seem to have suitable width 
or flatness for storing items. Such ledges may 
have paralleled the main roof support of the 
room, serving to support secondary roof and/or 
floor members. 

Vertical Ceiling Hole (Code 47) 

Of 59 examples of this feature (also 
known as Panowski holes), we recorded 58 at 
Tsankawi and one at Group A. The code was 
added during the Tsankawi recording, so it was 
not available at Groups F, I, and M, and the 
first half of A, but few if any vertical ceiling 
holes are likely to be present in those groups. 
The nature of the tuff at Tsankawi may in part 



186 CAVATE STRUCTURES 

Table 4.34. Vent Occurrence and Dimensions. 

A. Vent Occurrence by Shape and Group 



Shape 


Group A 


Group F 


Group I 


Group M 


Tsankawi 


Total 


Oval 





2 











2 


Trapezoidal 





1 











1 


Rectangular 





1 











1 


Cylindrical 


4 


4 


3 


3 


12 


26 


Truncated cone 


1 





1 








2 


Total 


5 


8 


4 


3 


12 


32 



Shape 



B. Chamber Location by Shape 

itWal tack Wall Left Wall Exterior Wall 



Total 



All Shapes 


2 


2 


6 


20 1 1 


32 


Dimension 


n 


C. Mean Dimensions for Round Shapes in Meters 

Standard 
Mean Deviation Minimum Maximum 


CV 


Diameter 
Depth 


23 
23 


0.18 
0.35 


0.068 
0.153 


0.04 0.33 
0.09 0.60 


38.6 
43.2 



Table 4.35. Groove Occurrence by Shape and Group. 



Shape 



Group A Group F Group I Group M Tsankawi 



Total 



Linear 


2 








1 


3 


6 


Rectangular 














1 


1 


Cylindrical 








1 








1 


Total 


2 





1 


1 


4 


8 



Note: Grooves were located on exterior, right, and back walls (two or three on each wall type), and one 
case was on a ceiling (a loom support). 



FEATURES 187 



Table 4.36. Wall Depression Occurrence and Dimensions. 

A. Wall Depression Occurrence by Shape and Group 



Shape 



Group A Group F Group I Group M Tsankawi 



Total 



Rectangular 
Rectangular solid 
Oval 

Cylindrical 
Hemispherical 
Conical 
Trapezoidal 
Circular portion 
Linear, irregular 



3 
7 
5 
1 
5 
1 
2 
1 







2 


2 


3 


1 


1 








1 


2 








3 


1 


2 


1 


1 


1 




















2 


1 






2 
2 
5 

8 

1 
1 




9 

14 
11 
3 
19 
4 
3 
2 
3 



Total 



25 



11 



19 



68 



Noncavate, grouped 





16 


3 







19 






B. Chamber Location by Shape 








Shape 


Right 
Wall 


Back 

Wall 


Left 
Wall 


Exterior Wall 
& Corners 


Back 
Corners 


Other 


Total 


Rectangular 


1 


18 


2 








2 


23 


Oval 


3 


5 


2 





1 





11 


Trapezoidal 





2 


1 











3 


Cylindrical 


3 

















3 


Hemispherical 


8 


7 


1 





1 


2 


19 


Conical 





1 


2 





1 





4 


Circular 
portion 














1 


1 


2 


Linear, 
irregular 


1 


2 














3 


Total 


16 


35 


8 





4 


5 


68 



188 



CAVATE STRUCTURES 



Table 4.36. (continued) 



C. Mean Dimensions by Shape in Meters 









Standard 








Shape 


n 


Mean 


Deviation 


Minimum 


Maximum 


CV 


Rectangular 


14 


0.03167 


0.0711 


0.0006 


0.2676 


224.6 


solid volume 














Hemispherical 


18 


0.01905 


0.0313 


0.0001 


0.1031 


164.1 


volume 














Oval area 


10 


0.1531 


0.1379 


0.0154 


0.4779 


90.1 


Rectangular 


7 


0.7004 


1.2910 


0.0391 


3.6100 


184.3 


area 














Height above 


23 


0.34 


0.218 


0.04 


0.86 


63.4 


floor" 















"Heights above floor include only chambers with less than 15 cm fill. 



explain their presence there, and it is unlikely 
that we overlooked their presence in the Frijoles 
sample recorded before the Tsankawi recording. 
The features consist of cylindrical vertical holes 
in the ceiling, sometimes extending to 
remarkable depths: the deepest recorded is 97 
cm. The average depth is 18 cm, and the 
majority are 5-10 cm deep. The holes are 
usually fairly small in diameter (mean of 4 cm, 
with 95 percent of cases less than 10 cm) and 
are often extremely regular. They sometimes 
look mechanically created, but many have 
smoke-blackened interiors and thus appear to be 
ancient. They seem generally to occur in 
groups; the occurrences range from 1 to 12, 
averaging 4 per cavate. The function of this 
feature type is unknown; they may be related to 
weaving, partitions, or construction. The 
vesicular nature of the upper stratum of 
Tsankawi tuff also opens the possibility that 
some are natural, though some are 
unquestionably artificial. 



Narrow Wall Incisions (Code 48) 

These features are also almost unique to 
Tsankawi, with the exception of a single case at 
Group A. The code was also added during the 
Tsankawi recording, so it was not available at 
Groups F, I, and M. As is true of vertical 
ceiling holes, codes were added as new features 
were observed, and we did not note these 
features before working at Tsankawi. 
Differences both in tuff type and in prehistoric 
activity, rather than recording differences, 
account for most of the distributions in this 
sample. If this feature were as common in the 
Frijoles cavates as it is at Tsankawi, we would 
have added it sooner. The incisions are usually 
vertical or close to it, and are 1-2 cm wide, 1-3 
cm deep, and up to 30 cm long (Figure 4.22). 
They appear to be incised by grinding. They 
generally occur in groups and were usually 
recorded as groups rather than individually; 26 
cases are recorded. They seem most often to be 



FEATURES 



189 



in smaller, unfinished rooms, though they are 
not confined to such rooms. Many more may be 
covered by plaster. It is possible that they were 
meant to help hold plaster, but this seems 
doubtful; they are more likely some artifact of 
construction, perhaps of cutting out blocks of 
tuff to be knocked free. Eleven of the cases 
recorded are on back walls, and eight are on 
right walls. 

Hand-or-Toe hold (Code 33) 

Generally these are not features one 
would expect to find inside rooms. The cases 
we observed either were cliff features (e.g., 
trails at Group A) or may result from 
remodeling. At Group F in the area of F-21, 



F-22, F-24, and F-25, for example, rooms 
appear to have extended quite far up one part of 
the cliff; probably after the rooms were gone 
(but perhaps before or even intermediate to 
building phases) a hand-and-toe hold route was 
apparently put across the same part of the cliff. 
The only individually recorded cases with rooms 
in our sample are four examples from Group M; 
two extramural sets were recorded as noncavates 
at Group A. A group of 13 holds was recorded 
at Group A, and 16 more were recorded as a 
group at Group F. 

Incised Dado (Code 29) 

We invented this feature type to include 
an observed phenomenon. In five cavates we 




Figure 4.22. Narrow wall incisions in TS-24. This feature type was observed almost exclusively at 
Tsankawi, and may. be the result ofcavate excavation or may have been to help plaster 
adhere to the wall. The vesicular nature of the top layer of the tuff at the Tsankawi 
group is visible here. 



190 CAVATE STRUCTURES 



recorded a distinct abraded band around 
substantial portions of the base of the chamber 
as defined by the top of the fill; in most cases 
the chambers containing this "feature" also had 
substantial amounts of fill. These bands are 
consistently 30-40 cm wide by about 6 cm deep, 
so that they seem intentional. We now believe, 
however, that incised dados are postoccupational 
damage, probably caused by large animals 
(probably sheep and goats, maybe burros or 
cattle) bedding down in chambers and rubbing 
against the walls. This interpretation is 
supported by the frequent association of dung 
with this "feature," and by its location at the top 
of the fill. 

Cliff Niche (Code 49) 

Cliff niches are distinctive features found 
only at Tsankawi. They consist of rectangular 
depressions in the tuff, mostly apparently outside 
of rooms, though quite possibly adjacent to 
rooftops. They look a great deal like fold-down 
metates, but hinges are lacking and gravity 
appears to be quite normal in the area. A group 
of five is located near TS-25 (Figure 4.23), 
where there is also a great deal of rock art. 
Indeed, they might even be considered rock art, 
as it is hard to impute a function to them. They 
are arranged so that three are next to one 
another and two more are widely spaced but 
close to the same level. Though they are several 
centimeters deep, the bases slope enough that it 
would be impossible to put anything in them. In 
the right light they look like doors, and they 
may have been designed to make the settlement 
look larger than it was from a distance (much as 
a cat tries to look big when frightened or a moth 
feigns to have large, scary eyes). The mean 
dimensions of the seven cliff niches we recorded 
are 0.62 m high by 0.37 m wide by 0.18 m 
deep. Mean door dimensions are fairly similar 
(0.70 m high by 0.56 m wide), and the cliff 
niche ranges fall well within the ranges for 
doors (Table 4.7). 



Another pair is located in the Tsankawi 
cavate group west of LA 50976, in the area of 
Lister's C-88-C-91. These niches are also 
located at the head of the rincon, on the caprock 
stratum, above cavates and probable masonry 
rooms. Of all the cliff niches observed, only 
one of these two is deep enough at the base (36 
cm) to hold something. 

Rock Art 

More often than not the rock art visible 
appears to be incomplete due to combinations of 
plaster deterioration, weathering, and vandalism. 
Usually rectangular boundaries of figures were 
estimated and measured for recording, though it 
was occasionally possible to measure actual 
features. June Crowder recorded the rock art 
separately (see her summary at the end of this 
chapter; see also appendix 4 at the end of this 
study), so the rock art entries on our recording 
forms were usually quite general. The rock art 
was categorized by subject matter and means of 
manufacture, with any form of painting being 
termed a pictograph and any form of incision or 
pecking a petroglyph. Petroglyphs thus include 
the fine-line scratching in plaster that seems to 
have been quite common; Chapman (1916; in 
Hewett 1938) found this style of rock art 
especially interesting (see also Schaafsma 
1980:285). Pictographs include any figures 
painted on the walls. We observed several 
colors of paint, including yellow, red, black, 
white, and green, and there are some figures 
that appear to have been done in a thin wash of 
plaster different in color from the wall plaster. 
Pictographs of all forms are less well preserved 
than incised or pecked rock art, harder to 
discern, and less frequent. 

The various rock art codes were applied 
as follows. Where two codes are present, both 
a petroglyph and a pictograph code were used. 
Geometric figures with definable geometric 
layout (codes 20, 22) include terraced figures, 



FEATURES 



191 




Figure 4.23. Four of the Jive cliff niches in the cap rock behind the Tsankawi cavates recorded by this 
project. Note that the bases of the niches are not flat enough to stand anything on. 



rectangles, circles, and crosses. Zoomorphic 
figures most commonly include parrots and other 
birds and Awanyus or horned (or plumed) 
serpents (codes 21, 23; Figures 4.24-4.28; 
Schaafsma 1980:255-288). Frequently it is clear 
that rock art exists, but it is either no longer a 
recognizable form or it may include several 
categories; these cases were called indeterminate 
(codes 24, 25). Early in the recording we 
observed handprints, which are a common 
prehistoric Pueblo form, and added a code for 
them (code 32); we did not, however, observe 
any more of them. The case we observed would 
be considered a pictograph since the prints are 
negative figures apparently done by spraying 
paint around an outspread hand. 
Anthropomorphs, on the other hand, proved to 



be fairly common (codes 34, 38); included in 
this category were fairly naturalistic figures 
(dancers and hunters) and supernatural figures 
(masks, katsinas, ogres). (See Figures 4.24- 
4.28.) The rock art tallied here does not include 
the examples observed at Groups A and I and at 
Tsankawi that are not associated with rooms 
(June Crowder did record these panels; see 
appendix 4). 

Since the majority of the rock art present 
consists of some small (and usually 
undeterminable) fraction of the original, we have 
not included measurements for the various 
panels, though measurements are present in the 
data base. 



192 CAVATE STRUCTURES 




Figure 4.24. Large katsina-like figures in TS-59. There is a terraced figure between them and an 
Awanyu line continues from them to the left. This chamber also contains a mask or 
katsina incised in the plaster next to a wall niche. This is the largest chamber recorded 
and that with the most features; it contains an array of loom features, floor features, and 
niches; viga holes and a low plaster "dado " are visible. The viga holes visible in this 
photo are of the variety that do not seem necessary for roofing; the presence of loom 
anchors and other upper loom supports in this room suggest that these "vigas * may have 
been involved with weaving. 



Not surprisingly, Group M and 
Tsankawi, the two groups with the most cavates 
have the preponderance of rock art (Table 4.37). 
Although they attract vandalism, cavates provide 
good conditions for preservation of rock art. 
Vandalism is most severe at Group A, increasing 
the chance that examples were overlooked there. 
Group M is notable for having the greatest 
variety of figures and media, including some 
elaborate polychrome panels, both within and 
outside the recorded sample. Group A is 
unusual in containing nearly equal numbers of 



painted and pecked or incised examples, while 
painted elements are infrequent at Group F. 
Though it is not apparent in the tables, the rock 
art at Tsankawi is notable for its larger scale. 
June Crowder summarizes her recording of rock 
art at the end of this chapter. 

Two preservation variables complicate 
the differences noted by Crowder between 
Tsankawi and Frijoles. First, the rimrock at 
Tsankawi is harder than any of the exposed 
cliffs in Frijoles-presumably that is why it is 



FEATURES 



193 




Figure 4.25. Large bird figures remain on either side of the door of TS-40, 
and a probable cloud motif is over the door. These figures are 
in the open on the very soft red tuff layer at Tsankawi, and must 
have been quite deep to have survived. 



1 


GF 
C23 


L._ ._ .. ... ... ... 1. 





Figure 4.26. Probable bird figure incised through some of the smoked tuff 
wall; F-23. 



194 CAVATE STRUCTURES 




Figure 4.27. Parallel zigzag lines, possibly Awanyus, in M-13; deeply chipped 
and abraded lines like this were seen elsewhere in Frijoles and 
also at Tsankawi. Patches of white paint are visible on the wall 
(this chamber also contained a handprint outlined in white). 
Vandalism to plaster is visible (scale =10 cm). 




Figure 4.28. Two masks or faces side by side in M-60. Notice how the figures 
appear to have been plastered over. This room also contains 
two metate rests (see Figure 4. 7). 



FEATURES 195 



Table 4.37. Rock Art Occurrence and Chamber Location. 

A. Rock Art Occurrence by Type and Group 



Rock Art Code 



Group A Group F Group I Group M Tsankawi 



Geometric 
petroglyph 

Zoomorphic 
petroglyph 

Indeterminate 
petroglyph 

Anthropomorphic 
petroglyph 



3 
9 
2 
6 



6 
11 

7 
13 



Total 



Geometric 


4 a 


4 









4 


6 


18 


petroglyph 


















Zoomorphic 
petroglyph 


3 a 


4 




6 




9 


12 


34 


Indeterminate 


3 


4 




4 




8 


12 


31 


petroglyph 


















Anthropomorphic 
petroglyph 


2 a 


1 




2 




4 


7 


16 


Geometric 


2 







4 




1 





7 


pictograph 


















Zoomorphic 
pictograph 















2 


3 


5 


Indeterminate 


8 


1 




1 




7 


5 


22 


pictograph 


















Anthropomorphic 
pictograph 















3 


1 


4 


Hand print 
pictograph 















3 





3 


Total 


22 


14 




17 




41 


46 


140 






B. Chamber Location by Type 












Back 


Left 




Exterior 








Type 


Right Wall 


Wall 


Wall 




Wall 


Ceiling 


Other 


Total 



16 
32 
14 
31 



196 CAVATE STRUCTURES 



Table 4.37. (continued) 



Type 



Back Left Exterior 

Right Wall Wall Wall Wall Ceiling Other Total 



Geometric 
pictograph 

Zoomorphic 
pictograph 

Indeterminate 
pictograph 

Anthropomorphic 
pictograph 

Hand print 
pictograph 







7 
4 
13 

1 
















1 















5 

22 
4 
3 



Total 



28 



62 



23 



134 



"Each includes 2 group-recorded examples. 

caprock. The remarkable extramural rock art at 
Tsankawi is confined to this caprock layer; the 
extramural rock art in the groups in Frijoles is 
now uniformly eroded and difficult to see, 
making comparisons difficult. Second, the 
plaster in Frijoles is more often in better 
condition than that at Tsankawi. Much of the 
Frijoles rock art we encountered was incised in 
the plaster, so that Frijoles and Tsankawi are 
again not quite comparable. Moreover, the rock 
art inside Frijoles chambers is not comparable to 
Frijoles cliff figures. These further cavate 
caveats do not imply disagreement with the 
differences noted by Crowder: there are several 
petroglyphs inside cavates at Tsankawi that are 
like nothing we saw in Frijoles, and it is likely 
that the extramural art was also different when 
it was all new. 

Steen (1979) gives the name Mortandad 
Style to bold figures incised through smoke 
blackening into the light tuff, creating a strong 
contrast between the white figure and the black 
background. He says this style occurs in a small 



area from Bayo to Ancho Canyons, the 
immediate vicinity of Tsankawi. Figures include 
Awanyus, Kokopellis, dancers, birds, and the 
Toltec sun god. Several cavates at LA 50976 
contain rock art that fits this definition. Steen 
considers this style to have been only briefly 
used in the late fourteenth century, though the 
ceramics found by James Maxon (1962) in a 
"cave kiva" containing Steen's prime examples 
of the style suggest to him a date of 1325. An 
important criterion of Steen's definition of the 
style seems to be that the incised lines are free 
of soot. As Maxon notes, "Apparently either 
there were few fires in the kiva after the figures 
were carved, or they were periodically cleaned, 
as the incised areas have little or no soot 
remaining in them" (Maxon 1962:2). It is 
unclear whether an expert in the Mortandad style 
would classify the sooted figures in TS-59 
(Figure 4.24) as belonging to that style. It may 
be that a later date for the style is correct, and 
that it represents use of the cavates after their 
primary occupation. Rock art executed in this 
fashion is either much rarer or absent in Frijoles 



FEATURES 



197 



cavates, perhaps lending some support to the 
notion of a territorial boundary between Frijoles 
and Tsankawi, at least in more recent times. 

Summary of Detailed Rock Art Study 

June Crowder 

The data recorded during this 
investigation are summarized in Table 4.38 and 
the tables in appendix 4 at the end of this study. 
The types of rock art and individual cavate 
contents are shown for both petroglyphs and 
pictographs for cavate Groups A, F, I, and M in 
Frijoles Canyon and for the selected group at 
Tsankawi (LA 50976). In addition, the cliff- 
face petroglyphs associated with the recorded 
groups are noted. Appendix 4 concludes with a 
brief correlation of the rock art recorded during 
this survey with that done by Chapman in 1916. 

Table 4.38 presents a breakdown of 
cavates containing rock art types as defined in 
Table 4.39. Though I made no formal compari- 
sons between this rock art and the examples 
described by Polly Schaafsma, our examples fall 
within the area and style she defines as the Tewa 
Division of the Rio Grande Style (Schaafsma 
1975, 1980). 

Many of the cavate walls and ceilings 
have multiple petroglyph designs. Because it is 
difficult to enumerate motifs and desirable to 
simplify the presentation of the recorded 
information, Table 4.38 and the group data in 
appendix 4 present the frequency of cavates 
containing each design type occurring within 
each cavate group, rather than the absolute 
frequency of each motif. The few pictographs 
found were for the most part unidentifiable. 
They are included in the summary table of each 
group as the number of cavates containing 
pictographs (appendix 4), with a brief 
description of each in the individual cavate table. 



The objective of this investigation was to 
record the type, number, and condition of 
petroglyphs and pictographs in selected areas of 
Frijoles Canyon and in a sample at Tsankawi. 
The photographs we took can serve as a baseline 
against which to measure the effects of natural 
erosional forces, pollution, and vandalism over 
time. For example, in this survey, I matched 23 
drawings from Frijoles Canyon done by Kenneth 
Chapman in 1916 (later published in Hewett's 
Pajarito Plateau and Its Ancient People [1938]) 
with the originals. Of these 23, 3 show severe 
deterioration and 3 show slight deterioration 
resulting from loss of wall plaster. 

There are three major differences be- 
tween the cavate wall drawings in Frijoles 
Canyon and those in Tsankawi. First, several 
two-horned serpents were found at Tsankawi but 
none in Frijoles Canyon. Second, Frijoles 
Canyon cavates contained many stylized parrots 
and other birds, while none was found at 
Tsankawi. Third, Tsankawi contained a larger 
number of anthropomorphic figures. 

In Frijoles Canyon, there was a notice- 
able difference between the art on the cavate 
walls and that on the cliff faces. The cavates 
contain a richer variety of art and more cere- 
monial figures than do the cliff faces. Art on 
the cliff faces at Tsankawi shows more variety 
than that at Frijoles Canyon and corresponds 
more closely to the cavate wall drawings. 

The total number of cavates that con- 
tained rock art was 46, of which 17 were in 
Group A, 7 in Group F, 4 in Group I, 6 in 
Group M, and 12 in LA 50976 (Tsankawi). In 
these cavates, 98 percent of the rock art was in- 
cised petroglyphs, 1.7 percent was abraded 
petroglyphs, and 0.3 percent pictographs. The 
cliff-face rock art from both Frijoles Canyon and 
Tsankawi consists only of petroglyphs. 



198 CAVATE STRUCTURES 



Table 4.38. Cavates Containing Rock Art by Group and Motif. 



Motif 



Group A Group F Group I Group M Tsankawi Total 



Abstract 


8 


4 


3 


4 


7 


26 


Geometric 


4 


3 


3 


3 


2 


15 


Geometric abstract 


1 


1 


2 








4 


Cross 


1 











1 


2 


Zigzags 


1 








1 





2 


Pictograph 


4 


1 


1 


3 


1 


10 


Anthropomorphs 








1 





6 


7 


Ceremonial figure 





2 


2 





2 


6 


Human figure 





1 





1 





2 


Mask 





2 


1 


3 


1 


7 


Hunter 











1 





1 


Flute player 














2 


2 


Stick figure 








1 


1 





2 


Realistic animal 








1 


1 





2 


Realistic bird 


1 











2 


3 


Realistic snake 


1 


1 








2 


4 


Serpent 














1 


1 


Two-horned serpent 














2 


2 


Serpent motif 














2 


2 


Snake motif 














2 


2 


Stylized bird 





1 





1 





2 


Stylized parrot 








1 








1 


Stylized insect 





1 











1 


Terrace 





1 








1 


2 



Total aboriginal motifs 21 



18 



16 



19 



34 



108 



Cavates with art 



17 



12 



46 



Modern graffiti 



11 



18 



FEATURES 199 



Table 4.39. Rock Art Nomenclature. 



Abstract 

Geometric 
Geometric abstract 
Terrace 
Zigzags 

Anthropomorphic 
Ceremonial figure 
Human figure 
Masks 
Quadruped 

Realistic animal 
Realistic bird 
Realistic snake 
Serpent 
Serpent motif 
Snake motif 
Stylized bird 
Stylized parrot 
Pictograph 
Modern graffiti 



Includes curvilinear forms (different circular patterns), straight and curved lines, 
miscellaneous forms (dots, etc.) 

Connected straight lines forming a design 

Design that has abstract elements and geometries connected 

Geometric with stepped sides 

Lines with sharp angles in alternate directions 

"Boxlike" drawing of the human figure 

Elaborately decorated (especially headwear) human figure 

Realistic rendering of human figure (modern?) 

Both simple and stylized masks 

All animal types not falling into the realistic category (cliff-face figures only in 
this sample) 

Drawings of identifiable animals 

Figures easily identified as birds 

Figures easily identified as snakes 

Extremely long and wide snake figure 

Extremely long and wide snake figure without a head 

Any double-line figure of a snake without a head 

Elaborate drawings of birds 

Elaborate drawings of birds with definite parrotlike beaks 

Painted figures 

Names initials, dates, etc. 



Preliminary Functional Analysis of Cavate 
Chambers 



Ultimately, one of the main things we 
want to know about cavates is their prehistoric 
function. Archaeological determination of 
function is ideally based on several categories of 
evidence, such as associated artifacts, debris 
from activities, well-grounded analogy to 
ethnographic groups, and architectural 
morphology. In the present study, we are 
limited almost entirely to the last category. 
Functional inferences must therefore be 
tentative; they are based on the study of feature 
co-occurrence, chamber size, and estimates of 
intensity of use. 

Feature Co-occurrence 

To examine feature co-occurrence, the 
cavate feature data set was used to generate a 
second data set through a series of merges from 
various data sets; each case of the new data set 
is a cavate (noncavates are excluded). The 
variables are chamber volume, coats of plaster, 
and the occurrence of ten feature categories (see 
Table 5.1). Variable values for the feature 
categories are the sum of all the occurrences of 
features in a particular category. Number of 
plaster coats was taken from the back wall of 
each chamber, or added manually from a 
consciously selected wall for cavates without 
back walls recorded. Volumes were included by 
merging information from the feature lines for 
chambers. The result was a data set with 175 



cases profiling the cavates recorded, though 
fewer cases had data for the plaster and volume 
variables (Table 5.1). 

The categories contain the following feature 
types: 

Walls: all masonry and natural walls 

Holes: indeterminate holes and possible latilla 
holes 

Beam features: viga holes, beam seats, and wall 
ledges 

Niches: wall niches and large floor-level niches 

Floor features: floor, firepit, floor burn, floor 
ridge, and floor pit 

Rock art: each recorded group of rock art (but 
not every figure) 

Doors: interior and exterior doors 

Vents: smokeholes and wall vents 

Other features: grooves, wall depressions, 
deflectors, narrow wall incisions 

Loom features: loom anchors and upper loom 
supports 



201 



202 CAVATE STRUCTURES 



Table 5.1. Summary of Chamber Attribute Occurrence. 



Attribute 



Mean 



Standard 
Deviation 



Minimum Maximum Median 



Number of 


175 


13.3 


11.51 


1 


64 


10 


features" 














Plaster coats 


165 


1.8 


1.97 





9 


1 


Volume 


156 


3.62 


3.10 


0.3 


18.6 


2.9 


Walls 


175 


3.0 


1.20 





6 


3 


Holes 


175 


3.3 


4.30 





24 


2 


Beam features 


175 


1.8 


2.76 





15 





Niches 


175 


1.1 


1.49 





8 


1 


Floor features 


175 


0.9 


1.31 





6 





Rock art 


175 


0.7 


1.95 





13 





panels 














Doors 


175 


0.6 


0.78 





4 





Vents 


175 


0.5 


0.98 





6 





Other features 


175 


0.5 


1.07 





9 





Loom features 


175 


0.3 


2.03 





24 






"Excluding walls. 



Some rare feature types and ceilings are not 
included. The minimum occurrence column in 
Table 5.1 shows that no feature type occurs in 
every cavate. Ten chambers have no walls 
recorded; these "chambers" are mostly very 
irregular or partial (a few at Group M were 
"remotely recorded" by means of a ranger 
calling down observations). These ten have 
been excluded from the rank-order correlations 
of occurrence (Table 5.2), as have floor 
features, since recording them depends on 
depositional accident. The same could be said 
for loom anchors, but they are included because 
loom supports are visible. 

A matrix of Spearman rank-order 
correlations (r^ for feature counts was generated 



for the chamber data set (Table 5.2). Given that 
many chambers will be tied at zero or one for 
some features, the rank-order correlations of 
feature occurrence cannot be considered reliable 
predictors of the likelihood of finding feature & 
if feature b_ is present. Siegel (1956:210) states 
that the effect of ties is to raise the value of r s , 
though there is relatively little impact, at least on 
small data sets. Although the nonparametric 
correlation is more appropriate to these data, 
experimental runs with Pearson Correlations 
give similar orderings of association. The 
correlations provide some ordering of 
associations among features, volume, and plaster 
coats, with the proviso that infrequent feature 
types tend to have lower maximum r s values. 
Since feature counts are retained, the 



FUNCTIONAL ANALYSIS 203 



Table 5.2. Co-occurrence of Feature Categories in Chambers. 



Feature 1 



Feature 2 



Neither Fl Only 



F2 0nly 



Both 



Holes 


Beam support 


50 


43 


16 


66 


Holes 


Niches 


49 


32 


17 


77 


Holes 


Vents 


60 


64 


6 


46 


Holes 


Other features 


60 


73 


6 


36 


Holes 


Rock art 


59 


74 


7 


38 


Holes 


Loom features 


65 


100 


1 


10 


Niches 


Rock art 


74 


59 


7 


35 


Niches 


Floor features 


58 


41 


23 


53 


Niches 


Vents 


70 


54 


11 


40 


Niches 


Loom features 


80 


85 


1 


9 


Beam supports 


Niches 


58 


23 


35 


59 


Beam supports 


Vents 


75 


49 


18 


33 


Beam supports 


Loom features 


90 


75 


3 


7 


Vents 


Floor features 


77 


22 


47 


29 


Rock art 


Floor features 


84 


15 


49 


27 


Rock art 


Loom features 


128 


37 


5 


5 


Floor features 


Loom features 


96 


69 


3 


7 


Other features 


Loom features 


127 


38 


6 


4 



correlations also give added dimension to the 
straight co-occurrence data (Table 5.3). Quite 
understandably, we find a fairly good association 
between chamber volume and the number of 
features present, but there is less likelihood that 
a larger chamber will have been plastered many 
times. Generally, and again predictably, the 
matrix shows that large numbers of features tend 
to occur together (see especially the feature 
number correlations in Table 5.2). 

Among the many correlations, several 



deserve note. The presence of niches seems to 
correlate well with holes, beams, volume, rock 
art, and plaster coats. There are somewhat 
surprising correlations between floor features 
and rock art, and floor features and plaster 
coats. These correlations may result in part 
from the fact that chambers showing floor 
features tend to be well-protected, well- 
preserved ones. The correlation between art and 
vents may be in part similarly explained, and the 
correlation between vents and doors is related to 
the presence of intact exterior walls. The 



204 CAVATE STRUCTURES 

Table 5.3. Spearman Rank-Order Correlations of Feature Category Co-occurrence. 

A. Plaster Coats, Number of Features, Walls, and Holes with Other Feature Types 



Feature 


Volume 


Plaster Coats 


Feature n 


Walls 


Holes 


Volume 
n 


1.000 
148 


0.378 
143 


0.631 
148 


0.247 
148 


0.485 
148 


Plaster coats 
n 




1.000 
160 


0.460 
160 


0.064** 
160 


0.386 
160 


Feature number 
n 






1.000 
165 


0.389 
165 


0.798 
165 


Walls 
n 








1.000 


0.275 
165 


Holes 
n 










1.000 
165 


Beam supports 
nt 


0.411 
148 


0.386 
160 


0.596 
165 


0.171* 
165 


0.373 
165 


Niches 


0.543 


0.467 


0.685 


0.216 


0.5220 


Vents 


0.445 


0.320 


0.556 


0.451 


0.403 


Rock art panels 


0.266 


0.384 


0.527 


0.111** 


0.300 


Floor features 


0.225 


0.472 


0.371 


-0.069** 


0.118** 


Loom features 


0.222 


0.134** 


0.274 


0.071** 


0.184 


Doors 


0.234 


0.093** 


0.426 


0.469 


0.309 


Other features 


0.226 


0.060** 


0.390 


0.180* 


0.246 



FUNCTIONAL ANALYSIS 205 



Table 5.3. (continued) 



B. Beams, Niches, Vents, Rock Art, and Floor Features with Loom and Door Features 



Feature 


Beams 


Niches 


Vents 


Rock Art 


Floor 


Beam supports 


1.000 


0.443 


0.225 


0.245 


0.208 


nt 


165 


165 


165 


165 


165 


Niches 




1.000 


0.335 


0.379 


0.289 


Vents 






1.000 


0.309 


0.200 


Rock art panels 








1.000 


0.343 


Loom features 


0.184 


0.273 


0.134** 


0.171* 


0.235 


Doors 


0.168* 


0.247 


0.480 


0.168 


-0.029** 


Other features 


0.117** 


0.194 


0.174* 


0.221 


0.189 



C. Loom Features and Doors with Other Features 



Feature 



Loom 



Doors 



Other 



Door 
Other 



0.046** 
0.100** 



1.000 



0.147** 



Note: The feature numbers used do not include walls as features; values for volumes are rounded to the 
nearest 0.1 m 3 for the calculations. 

♦Associated probability of no correlation greater than 0.01 (.01<p< .05). 

♦♦Associated probability of no correlation greater than 0.05. 

tThe ns for all succeeding lines in this block are the same as for this line. 



206 CAVATE STRUCTURES 



correlation between loom features and rock art 
is surprisingly low, an outcome surely 
influenced by the fact that TS-64 has 24 loom 
features (top rank) and only one rock art panel. 
Visibility and preservation again influence the 
relatively strong correlation between loom 
features and floor features; the correlation shows 
that loom features do occur with other floor 
features. 

Plastering and Smoking 

Plastering and smoking of cavates offer 
the best means for estimating duration and 
intensity of use of cavates. The majority of 
cavates exhibit wall plastering, but the finish of 
the coats and the number applied are quite 
variable (see Tables 4.9, 4.10). Although it is 
not possible to determine the interval between 
plasterings, the presence of several coats of 
plaster seems a good indication of maintenance 
of the room; when numerous coats are present, 
they are usually smoked, which in turn shows 
that the maintenance was occasioned by use of 
the room. If this reasoning is valid, we may ask 
such questions as which rooms are maintained, 
whether apparently long-used rooms contain 
more features, and whether large rooms are 
more frequently maintained than smaller ones. 

The maximum number of plaster coats 
recorded is 10, with the most on a back wall 9 
(and thus transferred to the chamber data set). 
Table 5.4 shows the co-occurrence of coats of 
plaster and number of features, excluding walls. 

This tabulation gives dimension to the 
Spearman correlation value of 0.46 for plaster 
coats and number of features. That is, chambers 
with very few features clearly tend to have little 
plaster, and the rooms that have the most 
features tend to show at least some replastering. 
Still, some rooms have multiple replasterings 
and relatively few features, and the rooms with 
the most features are not those with the most 
plaster coats. This finding may be related to 



personal variation in level and frequency of 
maintenance. It also seems to indicate that the 
rooms that were used the longest (or maintained 
the most frequently) were not those in which the 
most activities took place. Because the 
correlation between coats of plaster and chamber 
volume is even lower (r s =0.38), it is tempting 
to suggest a division between rooms of everyday 
use (smaller, more often replastered) and 
special-use rooms (larger. m ore features, less 
often plastered). This interpretation is 
countered, however, by the fact that all the 
largest chambers are located at Tsankawi, where 
plaster is either less frequent or less well 
preserved. 

These data give some indication of the 
periodicity and perhaps reason for replastering. 
On all walls with many coats of plaster, at least 
some of the coats are smoked, but all the coats 
are smoked in only a few cases (Table 5.5). 
Often, we found that the most recent coat looked 
quite clean, which may say something about 
abandonment practice or may indicate that that 
was the desired state (barring some unfortunate, 
unknown recent attempt at "restoration"). In the 
150 walls of rooms designated as storage rooms, 
only 21 percent were smoked (the maximum 
number of coats in these rooms is two). In 
rooms assigned to the habitation group, on the 
other hand, 71 percent of 303 walls showed 
some smoking, while 86 percent of 29 "kiva" 
walls showed smoking. 

Cluster Analysis 

Each chamber recorded on a cavate form 
was assigned to a functional category by the 
recorders. These categories include habitation, 
storage, "kiva," and unknown. We based these 
assignments on the presence of various feature 
types, chamber size, and wall treatment. Merely 
tabulating functional categories by these features 
only shows how much latitude the recorders 
allowed for each category (that is, it quantifies 
their mental templates for each function). The 



FUNCTIONAL ANALYSIS 207 



Table 5.4. Plaster Coat-Feature Number Co-occurrence. 



Number of Features Coats 1 Coat 2 Coats 3-5 Coats 6-9 Coats Total Percent 



1-2 

3-4 

5-6 

7-9 

10-12 

13-16 

17-20 

21-25 

26-41 

43-64 



9 
8 
8 
5 
6 
4 
1 
3 
2 




6 

8 

10 

7 
6 

5 
4 
4 

1 
1 



3 


2 
3 
1 
2 
3 
2 
3 
2 







2 

2 
1 
3 
4 




18 
17 
22 
19 
16 
18 
16 
18 
15 
6 



10.9 
10.3 
13.3 
11.5 

9.7 
10.9 

9.7 
10.9 

9.1 

3.6 



Total 



46 



52 



21 



34 



12 



165 



99.9 



analysis can be taken one step further by 
performing a quantitative classification analysis 
based on measurements and the presence of 
features and plaster. Such an analysis allows us 
to see if this more objective approach finds 
different categories and whether the subjective 
classifications are to any degree independently 
created. 

The co-occurrence information is 
suggestive (Table 5.2), but it rapidly becomes 
unwieldy because of the very large number of 
possible combinations. In an effort to place 
chambers in groups based on feature occurrence, 
several cluster analyses were performed using 
the SAS FASTCLUS procedure. The following 
variables were included: volume, plaster coats, 
number of holes, number of niches, number of 
beam supports, and number of rock art panels. 
Infrequent and inconsistently observable 
categories, such as loom and floor features, 
doors, and some other features, were not 



included, and completeness and fill filters were 
also applied, resulting in a group of 127 rooms. 

Analyses were done requesting 6 and 10 
clusters. The sizes of the clusters formed are 
quite different. In the 6-cluster version, one 
cluster contains 78 of the cases; the 10-cluster 
version breaks this cluster into two clusters of 
57 and 25 (some cases change cluster 
membership when new clusters are present). 
Though it generates several clusters with only a 
few members, the 10-cluster analysis is 
presented because variability in function is a 
focal question. Even with 10 clusters, cases that 
are far from the cluster seed in large clusters are 
intuitively unlike the cluster profile (see 
appendix 5). 

The clusters formed by this procedure on 
the whole make intuitive sense, though a few 
placements seem odd based on a knowledge of 
the room and other preconceptions. Once again, 



208 CAVATE STRUCTURES 



Table 5.5. Smoking of Plaster Coats. 











Number of Smoked Coats 






Total Plaster Coats 


No Smoke 


1 


2 


3 


4 


5 


6 


7 


9 


No coats 


126 


55" 
















One coat 


93 


48 
















Two coats 


19 


40 


6 














Three coats 


4 


25 


16 


3 












Four coats 


2 


10 


12 


5 


1 










Five coats 


1 


2 


4 


17 


3 











Six coats 








3 


3 


7 












Seven coats 








2 





3 


1 





1 




Eight coats 




















2 







Nine coats 























1 


1 


Ten coats 























2 





Total 


245 


180 


43 


28 


14 


1 


2 


4 


1 



"One "coat" of smoking on no coats of plaster indicates a smoked, unplastered surface. 



it is important to recognize that the program 
"knows" only the variables given it and that 
each of those variables is given equal weight. 
That is, as far as the program is concerned a 
chamber's volume is just as important to its 
placement as the number of indeterminate holes, 
and for this analysis it did not consider the 
presence of loom features or the scope of the 
rock art (merely the abundance of the panels). 
As is apparent in the comparison of cluster 
placement to assigned function, the clusters also 
indicate that subdivision of the assigned 
categories is possible and that the categories do 
overlap. Cluster membership for individual 
cases is given along with volume, plaster coats, 
and number of features in appendix 5. 

The cluster means give a fairly good 
idea of the characteristics of each cluster 



(Table 5.6). Cluster 6 chambers are relatively 
low volume, have few plaster coats, and few 
features. They thus correspond fairly well to 
what we called storage rooms, though some 
chambers we thought to be habitation rooms are 
included. This cluster also accounts for most of 
the chambers placed in the unknown function 
category. This cluster is found in all groups, 
more or less proportionally to the size of the 
sample included from each group. Therefore, it 
may be suggested that such low-activity rooms 
were a part of the complement of all cavate 
settlements, and that they were likely to have 
been for storage. 

Four clusters (1, 2, 3, 7) contain rooms 
only from Tsankawi. As noted earlier, many 
Tsankawi rooms are considerably larger than 
those recorded in Frijoles, and large size is an 



FUNCTIONAL ANALYSIS 209 



Table 5.6. Means, Membership, and Correspondence to Assigned Function for Chamber Ouster 
Analysis. 



Cluster 



n 



A. Ouster Means for Chambers 

Volume Plaster Holes Art Niches Beams 



1 


2 


5.48 


1.5 


22.5 


2.5 


2.0 


4.0 


2 


5 


10.92 


1.6 


10.0 


1.0 


4.0 


2.6 


3 


3 


17.04 


3.0 


13.7 


7.7 


5.3 


8.7 


4 


4 


5.60 


5.8 


6.2 


0.8 


2.8 


11.8 


5 


13 


5.24 


4.9 


9.1 


1.0 


2.3 


2.3 


6 


57 


2.16 


1.1 


0.5 


0.2 


0.5 


0.3 


7 


2 


15.15 


1.0 


5.5 


0.0 


4.0 


5.5 


8 


2 


3.99 


3.0 


3.0 


12.0 


1.5 


1.5 


9 


25 


3.01 


1.7 


5.6 


0.6 


1.3 


1.0 


10 


15 


4.30 


1.7 


2.3 


1.2 


1.4 


5.5 



B. Chamber Ouster Membership by Group 



Cluster 



Group A Group F Group I 



up M 


Tsankawi 





2 





5 





3 


2 





1 


2 


12 


22 





2 


1 





1 


11 


8 


4 



Total 



1 
2 
3 
4 
5 
6 
7 
8 
9 
10 







2 
5 
9 


3 
1 







5 
4 


5 









10 


1 

5 
2 



2 

5 

3 

4 

13 

57 

2 

2 

25 

15 



Total 



20 



14 



18 



25 



51 



128 



210 CAVATE STRUCTURES 



Table 5.6. (continued) 



Cluster 



C. Chamber Cluster Occurrence by Assigned Function 

Habitation Storage Kiva Unknown 



Total 



1 
2 
3 
4 
5 
6 
7 
8 
9 
10 



1 





1 





2 


4 





1 





5 


1 





2 





3 


3 





1 





4 


13 











13 


15 


35 





7 


57 


2 











2 








2 





2 


17 


5 





2 


24 


13 


1 





1 


15 



Total 



69 



41 



10 



127 



important aspect in three of these Tsankawi-only 
clusters. Clusters 3 and 7 contain the largest 
rooms in the sample; the difference between 
them is that Cluster 3 rooms not only are the 
largest but also have some of the highest feature 
counts (in addition they have multiple plaster 
coats, while the cluster 7 rooms do not). All 
attributes suggest that TS-20, placed in cluster 3 
and called a habitation room by the field 
recorders, should be considered in the same 
category ("kiva," if you will) as the other rooms 
in cluster 3 (TS-59 and TS-66). In fact, the 
recorders noted the presence of upper loom 
supports and suggested that the room was 
possibly used as a "kiva." The frequency of 
rock art is very high in the chambers in cluster 
3. The two members of cluster 1 are well 
within the size range for Frijoles chambers; they 
differ, however, by having fewer coats of plaster 
than other rooms with an equal number of 



features and by containing an inordinate number 
of hole features. Cluster 2 may be characterized 
as large rooms containing many features but 
little rock art, few coats of plaster, and many 
niches; TS-64, the single chamber containing the 
most weaving features of any recorded and 
called a "kiva," is placed in this group. This 
cluster is easy to understand as a group of large 
chambers with intermediate numbers of features, 
especially because loom features were not 
included in the analysis. 

Clusters 4 and 5 might be considered 
"typical" high-activity Frijoles rooms. Volumes 
are relatively high for Frijoles, and the chambers 
have many features (but infrequent rock art) and 
many coats of plaster. The difference between 
the two is primarily a high frequency of beam- 
supporting features in the Cluster 4 rooms. A 
couple of Tsankawi rooms fall within this 



FUNCTIONAL ANALYSIS 211 



constellation of attributes. Cluster 8 contains 
two Frijoles "kivas" that are in fact quite 
similar, M-13 and 1-19. They are both entirely 
enclosed, relatively small chambers with much 
rock art (this cluster has the highest mean 
frequency of rock art) and abundant other 
features. 1-19 also contains loom features and 
floor features not used in the analysis. 

Cluster 9 is a good deal like cluster 6, 
the storage cluster, though its members are 
somewhat larger and have more abundant 
features of all types; the majority of this cluster 
was classified as habitation rooms rather than 
storage rooms. This cluster may indicate the 
existence of a group of rooms exhibiting less 
evidence for intensive activity but still differing 



from small, featureless rooms. Cluster 10 seems 
to be yet another small increment in apparent 
activity; rock art, niches, and especially beam 
supports are more frequent in this group than in 
clusters 6 and 9. 

The chamber cluster analysis is really 
only a starting point for attempting to identify 
functional variability in cavate chambers. The 
criteria could be refined and the combinations 
elaborated through the use of more feature types 
with tighter definitions. Much more information 
is needed on the functional significance of the 
features themselves. Still, the differences and 
similarities in the classifications do suggest that 
this may be a useful approach to examination of 
larger samples. 



Interpretation and Conclusion 



The field recorders on this project were 
asked to attempt to assign one of three general 
functions to each cavate: habitation, storage, or 
"kiva" (Table 4.3B). These assignments are of 
course subjective, but they are also in a sense 
multivariate, because a human observer can 
form an impression based on a great many 
attributes that are difficult to quantify for 
supposedly objective classification by machine. 
The distributions of features within the field 
function groups are instructive (Table 6.1). 

This tabulation shows clearly the 
existence of a fairly large group of cavates that 
have small numbers of features and that do not 
look "lived in." There is also a much smaller 
number of cavates that show a great deal of 
evidence for activity, perhaps beyond just 
habitation. 

In 1952 Watson Smith published a 
chapter called "When Is a Kiva?" in which he 
neatly demonstrated that what archaeologists 
called kivas could not be defined by any single 
feature (Smith 1952:154-165; see also Thompson 
1990). Moreover, he showed that there were 
large areas of overlap between regular rooms 
and kivas in terms of size and feature 
occurrence. Having concluded that even within 
areas there is a great deal of architectural 
variability in kivas, Smith made several 
observations that remain useful here: 



It seems to me that the most convincing 
determinant in most cases lies not in the 
specific features of any particular room 
taken in the absolute, but rather in that 
room's relationship to other rooms in 
the architectural unit of which it was 
part. . . . 

It is my feeling that in a large number 
of cases a given room may have served 
for both secular and ceremonial uses, 
and that in most of those cases there is 
no way whatever of certainly 
determining the fact. On the other 
hand, there are many cases in which the 
relationships of a particular room within 
its architectural complex, its difference 
in shape or size from other associated 
rooms, and it positional relation to 
them, will be of greater significance in 
its identification than any or all of its 
internal features as such. (Smith 
1952:161-162) 

Peckham (1979) further elaborates the lack of 
consistency in ceremonial structures for the Rio 
Grande. 

Given the constraints imposed by "the 
medium" (cliffs) on cavate placement, form, and 
size, the problems reviewed by Smith are 
compounded for cavates. The use of the term 
kiva has three aspects: architectural, functional, 



213 



214 CAVATE STRUCTURES 



Table 6.1. Frequencies of Features by Assigned Function. 
Number of Features H 



ation 


Storage 


Kiva 


Unknown 


Total 


8 


34 





8 


50 


14 


13 





7 


34 


16 


4 





1 


21 


20 





2 





22 


27 


3 


1 





31 


3 





1 





4 


3 





4 





7 



1-5 

6-10 

11-15 

16-20 

21-30 

31-40 

41-64 



Total 



91 



54 



16 



169 



Note: Excludes the wall category; some rare feature types are not included. 



and semantic. In addition to the architectural 
variability noted for "kivas" in general, surely 
there was also considerable functional 
variability, as adumbrated by the chamber 
cluster and other analyses (see chapter 5). 
Semantically, the connotations of the term itself 
may or may not be warranted. 

Stephen Lekson (1988) has persuasively 
argued that the term Java as applied to 
prehistoric Pueblo archaeology and also the 
ceremonial connotations that it bears are more 
an artifact of political and practical aspirations of 
early twentieth-century archaeologists than the 
result of good archaeological evidence for the 
ceremonial use of kivas. Hewett was a prime 
proponent of early "kivas," and it was probably 
he who started the tradition of cavate kivas 
(Hewett 1909b:655-663). It seems extremely 
likely that the term is misleading as to the 
function of larger cavates with multiple features 
and rock art. Even the very largest cavates 
would not hold a gathering of any size. The 
many features of some larger examples are at 
least as likely to be the result of use primarily as 
a habitation room as they are to be the result of 
use as a ceremonial chamber. Among the 



attributes that prompted Hewett to call some 
cavate rooms kivas are a "ceremonial opening," 
murals, and rows of six or seven loops, which 
we call loom anchors, in the floor. He called 
the loops "a feature to be found on the floor of 
nearly every kiva that has been examined in this 
region" (1909b:660). These three features are 
present in all three kivas described by Hewett. 
Other features present at some of the kivas he 
describes are partial subterraneity, densely 
smoked walls, a plaster dado, a firepit, a sipapu, 
a posthole for a sun observation post, and an 
altar (or at least the place for one). While one 
of the three is the "largest cave kiva that has 
been found, another is "about 8 feet by 8 feet in 
dimension" (about 2.4 x 2.4 m; 1909b:658, 
655). 

Loom anchors are associated with kivas 
in other prehistoric Pueblo areas (see, e.g., 
Smith 1952:159; Magers 1986:270-271) and in 
the Rio Grande (Peckham 1979), but this means 
only that similar activities took place in some 
cavates and some other prehistoric Pueblo 
"round rooms," not that such rooms were 
strictly ceremonial in nature. The practice of 
weaving in Hopi kivas, however, does 



CONCLUSION 215 



strengthen the link between prehistoric and 
ethnographic kivas. It also serves as a caution 
that strict partitioning of daily activity from the 
ceremonial realm is not a realistic perspective 
for pueblos, prehistoric or modern (see also 
Steen 1977:17). Loom anchors in cavates tend 
to be found in chambers with many features, 
some of which are rock art panels. During 
fieldwork, our impression was that if one or two 
looms were set up in a chamber there would be 
room for little else, though perhaps looms could 
be taken down so that chambers could be used 
for other functions. Historically, cotton weaving 
took place mostly in the winter at Hopi (Kent 
1983b:27), which raises interesting if hard-to- 
verify possibilities for the seasonality of use of 
cavates. Many of the old photographs of Pueblo 
weaving published by Kent (1983b), however, 
show weavers working at outside looms, 
indicating that interior weaving was probably 
only a portion of the industry. 

Some cavates, then, were used for 
multiple activities and were probably somewhat 
similar to earlier prehistoric Pueblo kivas found 
to the west. Any conclusion about the 
ceremonial use of these rooms by a specific 
social group, such as a clan, would be extremely 
speculative, and one must keep in mind the 
severe space limitations of these rooms. To 
identify small cavates used for individual 
meditation cells or personal kivas (see Steen 
1977:14) would be virtually impossible. 

Smith's point that some rooms are 
distinctive in some way-shape, size, placement, 
features-is well taken with regard to cavates. 
Of the eight chambers that were considered to fit 
the "kiva" pattern, four are in the Tsankawi 
sample, and three of those are nearly contiguous 
(TS-59, TS-63, TS-66). This concentration of 
large, high-activity rooms suggests that most 
cliff suites with cavates probably had such a 
room; at Tsankawi, where a higher percentage 
of rooms were cavates rather than partial cavates 
or mostly masonry, more such rooms are 
visible. But were they "kivas"? In the sense of 



a feature where many people gathered for 
ceremonies, probably not. On the basis of little 
evidence, I would suggest that function probably 
took place at the various large masonry 
structures that are near most dense groups of 
cavates. In the sense of a room where ritual 
observance and preparation sometimes took 
place, doubtless these chambers were kivas. 
Clearly, after all these years, we still need to 
know not only when is a kiva but what is a kiva. 
At least archaeologically, the term now covers 
too many sins. 

Another function suggested for cavates 
is that of field house. Preucel has studied 
Pajarito field houses in depth (Preucel 1985, 
1986a,b), and has suggested that the cavates in 
our study could have performed that function (R. 
Preucel, personal communication, 1988). 
Dietrich Fliedner (1975) mentions "small caves" 
that have been modified in the vicinity of 
Unshagi on the Jemez River and suggests that 
these features too may have been used by small 
groups tending fields. The smallest of the 
groups in the present study was part of a site of 
at least 50 rooms, and all are in close proximity 
to further aggregations of population. This 
context does not suggest structures intended 
primarily for tending nearby fields, but much 
more isolated cavate rooms are present around 
the plateau in locations that do suggest this 
function. Comparison of such locationally 
distinct cases to the metric and feature profiles 
generated in the present study will help show 
whether they conform to general cavate 
construction and layout or whether they have 
distinctive attributes that pertain to their 
function. 

Whatever the function of individual 
chambers, the loom features found in cavates 
indicate that weaving was probably a very 
important, perhaps even major activity on the 
Pajarito Plateau. Detection of the presence of 
loom features is subject to several variables of 
preservation and deposition, and there is little 
doubt that rooms for which we did not record 



216 CAVATE STRUCTURES 



loom features do in fact contain them. Even so, 
10 of 175 (6 percent) chambers recorded had 
loom features observed. Some of the 175 are 
not in any condition to retain evidence of loom 
features, many others are far too small to 
contain a loom, and the floors of still others are 
not visible without excavation. If we grant that 
about 100 of these rooms were used for living 
activities, the percentage is even higher. The 
Pajarito Plateau seems an unlikely place to grow 
cotton, though the Spanish found the Pueblos 
growing it as far north as the mouth of the Rio 
Chama (Kent 1983b:5). This high frequency of 
loom features raises a host of questions: Was 
cotton a major crop? Was raw cotton a major 
import? Were the cavate inhabitants weaving 
some other fiber? Was weaving an important 
economic hedge for plateau inhabitants, an 
entree into the broader fifteenth-century 
economy? Whatever the answers to these 
questions, the apparent importance of weaving at 
these sites casts a different light on the historic 
association of weaving with the Hopi area (see 
Kent 1983a,b). 

Cavate Use and Relationship to Large 
Surface Pueblos 

All the cavate groups included in this 
study are to some degree spatially associated 
with large, free-standing pueblos, similar to the 
pattern observed elsewhere on the Pajarito 
Plateau (Preucel 1987). Group A is perhaps the 
farthest removed from such a site, but it is only 
1 km from Tyuonyi, and there is an apparently 
sizable site on the south side of the Rito between 
Ceremonial Cave and Group A. Group F is the 
closest to two large sites: Tyuonyi and the 
unexcavated house (Tyuonyi Annex - LA 60550) 
just to its east. Group I is close to these sites as 
well, Group M is above Rainbow House, and 
the LA 50976 cavates at Tsankawi are just 
below Tsankawi Pueblo (LA 211). The broad 
chronological information available (Tables 2.1, 
2.3) indicates that the two site types were used 
contemporaneously. The relationship between 



these two types is of great interest, but 
knowledge of the actual function of either type 
is at present so sketchy that discussion of the 
relationship must take the form of a listing of 
possibilities and the data available to support 
them. In assessing these relationships, we 
should consider that although the two site types 
make a neat archaeological dichotomy in terms 
of location and present appearance, this 
difference was probably less during occupation. 
The cavate groups studied in Frijoles were all 
substantial masonry structures that also happened 
to have back walls and some rooms that were 
part of the cliff. The plan of Long House (also 
known as Group D) with its tiers of masonry 
rooms and multiple stories is a good example; in 
all probability Group A was a site comparable to 
Long House in size and layout. 

Bandelier (1971) envisioned that the two 
types of structure were contemporaneous, that 
the people living in both were members of the 
same tribal unit but belonged to different clans, 
and that those in the cavates were somehow 
slightly less reputable. His vision accepts (or 
helped create) the received wisdom that large 
masonry structures were places where people 
lived year-round, and there does seem to be a 
great deal of overlap in feature type between 
cavates and free-standing pueblos. Both 
Bandelier and Hewett conceived of the large 
pueblos as "community" structures, and that 
concept has some appeal: they have a visibly 
organized layout as opposed to the seemingly 
more accretionary cavate/cliff sites. McKenna's 
ceramic analysis of limited cavate surface 
samples suggests that the cavates may have 
higher frequencies of culinary wares than the 
larger pueblos, which may be interpreted as 
indicating a greater domestic use of the cavates. 

Preucel (1987) has proposed using a 
succession model for occupation of the Pajarito 
Plateau. The latter two stages of his model 
(equating to the Late Coalition and the Early 
Classic) involve increased population 
concentration. Densely settled groups of cavates 



CONCLUSION 217 



are an effective means of aggregating population 
with minimal loss of productive land either on 
mesa tops or in canyon bottoms. Once again, 
the complementarity of free-standing masonry 
pueblos and cavate-using "talus pueblos" seems 
important. The type and degree of relationship 
is of critical importance to analyses such as 
Preucel's: should cavate floor areas be included 
with those of adjacent free-standing pueblos, or 
should the two types of structures be considered 
separate "sites"? 

Conclusion 

Any archaeological project has 
aspirations on several planes. On a mechanical 
and methodological level, this study presents a 
system for recording cavate rooms and features 
in a relatively rapid, reproducible fashion. 
Details of definitions and technique could 
certainly be refined; the metric analyses of the 
data collected as well as the experience of using 
the system on a large body of data point toward 
some of those refinements. The features studied 
were recorded in sufficient detail using enough 
different media that for these groups cultural 
resource managers have the information required 
for monitoring and preserving an important and 
rich archaeological resource. 

The studies of occurrence, co- 
occurrence, and metric variability will allow 
comparison with features and rooms-whether in 
cavates or not-recorded elsewhere on the 
Pajarito Plateau, to isolate morphological groups 
within general feature categories. Here the level 
of aspiration rises to determining function and, 
beyond that, better understanding culture history 
and ultimately cultural process. As is frequently 
the case, these aspirations are frustrated to 
varying degrees. The experiments with cluster 
analysis suggest that it is a potentially useful 
means of identifying classes of features or rooms 
that may have functional significance. The 
results of the cluster analysis also corroborate 
the indications from the other analyses that the 



broad storage-habitation classification has many 
subcategories of use and that it is probably 
possible to identify them. Though cavates 
provide a great deal of "cheap" data, one must 
remember that they are often only parts of sites. 
To fully understand those sites requires seeking 
information from the whole. 

Some questions remain largely 
unanswered, such as the question of a cultural 
boundary separating the Tsankawi area from the 
Frijoles area during the occupation of the 
cavates. We found some hints of such a 
boundary: features and rock art occurring only 
at Tsankawi, chamber size, and ceramic 
distributions. The indications are far from clear- 
cut, however, and those relatively small 
differences are seen against a backdrop of great 
similarity in a majority of other attributes. In 
that sense, though, the differences resemble 
those between the linguistically different 
ethnographic groups that claim these two areas 
(Tewa and Keres). Other issues, such as the 
relationships among the Frijoles cavate groups 
or between the cavates and the large free- 
standing pueblos, or the impetus to occupy and 
abandon cavates, remain speculative. Further 
research is necessary to shed light on these 
questions. 

This project is only a starting point for 
understanding and recording cavates on the 
Pajarito Plateau. The sample is sizable, but it is 
a very small percentage of the cavates that are 
"out mere." The study shows that cavates vary 
considerably, and, once subdivisions are made, 
the small sample size relative to the variability 
becomes apparent. All of the cavates included 
are from major clusters, and there are isolated 
ones which would make interesting comparative 
material. Excavation data from cavates would 
add important missing dimensions to this study, 
but, as shown by the nearly limitless numbers of 
potential combinations made apparent by this 
study (only a fraction of which have been 
explored here), cavates are remarkable for the 



218 CAVATE STRUCTURES 



amount of information available without 
excavation. 

The information available in cavates 
provides a perspective on prehistoric architecture 
that is rarely available from excavation. 
Correctly and necessarily, the focus of 
excavation data is the floor; more often than not, 
information on the rest of the room stops a few 
centimeters above the floor. In a cavate room, 
on the other hand, full, constructed walls and 
ceilings may still be observed, and the 
preservation is frequently excellent. In cases 
where floors have not been exposed to human 
and animal wear, preservation of organic 
materials and feature components is also likely 
to be very good. The numbers and variety of 
features on walls and ceilings catalogued by this 



study show how incomplete the floor perspective 
on prehistoric life can be. The high state of 
preservation in many cavates does not, however, 
mean that the information still available in 
cavates (with or without excavation) will last 
forever. Cavates on the Pajarito and elsewhere 
owe their existence to the softness of the rock 
into which they are dug. This means that 
features are continually degrading, especially 
where they are often visited. Archaeologists and 
managers should therefore take every 
opportunity to catch as much of the perspective 
provided by cavates as possible before it 
weathers away or is vandalized. There is much 
to be done, but there is solace in the thought that 
a great deal of information can be captured at a 
relatively low cost. 



Appendices 



219 



Appendix 1: Forms, Coding, and Materials Collected 

Forms Used in Recording, 1986 

Cavate Room Record 



Record number: 

Date: 

Site number: 

Photo key: 

Location: V* V*, Sec. T R 

Elevation: 

Exposure/orientation of exterior opening: ° TN 

Height above PGS of base of opening: m 

Fill: type approximate depth 

Estimate of completeness of room: 

Evidence for construction: 

Estimate of natural vs. excavated space: 

Masonry: 

Cavate rooms in group up-canyon 

Stability: 

Unusual tuff characteristics: 

Nonhuman use: 

Function: 

Room notes: 



Recorder: 
Cavate group: 
Quad: 
UTM: 



m 



Cavate no. 



N 



down-canyon 



Doors: 

No. Type # Location 
Part/Az. 



Features 

Shape Measurements 

a b c w th. 



Geo To Photo Feat 
f room f 



Door notes: 



221 



222 CAVATE STRUCTURES 



Cavate 



Page- 



Record No. 



Chamber: 

Locat. Shape Measurements Geo 
No. Type # Pt./Az abed f 



Plaster Damage 
n sm-n clr ht h n 



Fea 



Notes 
f 



/ ° 


/ ° 


/ ° 


/ ° 


/ ° 


/ ° 


/ ° 


/ ° 


Notes for no.: 




i 







Features 

No. Type # Location 
Part/Az. 



Shape Measurements Geo Height Photo Fea 
a b c deep f /floor f 



APPENDIX 1 223 



Cavate 



_Page_ 



Record No. 



Features: 

No. Type # Location 



Part/Az. 



Shape Measurements Geo Height Photo Fea Notes 
a b c deep f /floor f 



o 
o 
o 
o 
o 
o 
o 
o 
o 
o 
o 



Feature notes for no.: 



224 CAVATE STRUCTURES 



Noncavate Cliff Feature Record 



Date: 

Site number: 
Cavate group: 
Photo key: 
Location: V4 



14, Sec. 



Feature type: 

Human damage: 

Connected with cavate: 

Evidence for construction: 

Masonry: 

Stability: 



R 



Recorder: 
Record number: 
Room number: 
Quad: 

Elevation 

Natural damage: 



No. feature/part shape 



w 



Dimensions: 
h depth 



gf photo ff 



No. Type 



shape 



Features: 
size range 



gf photo 



ff 



Notes: 



APPENDIX 1 225 



Procedures and Codes for Recording 
Procedures and Codes for the Cavate Form 



Feature here is an inclusive term that includes chamber portions, doors, walls, firepits, 
and loom anchors. There are certain classes of information which do not apply to certain sets of 
features (firepits do not go to other rooms, and a wall does not need a precise location). 
Remember that codes can be added for things not covered in this list; added codes are 
documented as to definition and date of addition in a later part of this appendix. 

All measurements are in meters, with the number of decimal places reflecting the 
accuracy of the measurement. 

Record no. 

Every cavate record receives a record number which serves to tie together basic 
information and feature information. This is a very good number to get right on all the forms. 

Site 

Laboratory of Anthropology numbers (LA ) were assigned. Surprisingly, only a few 

of the cavate groups had LA numbers: Group D or Long House=LA 13665; Group E or 
Hewett's Sun and Snake Houses= LA 13664; Tsankawi itself=LA 211 (cavates around 
Tsankawi had no LA numbers). The following numbers were used in the field in 1986: Upper 
Group M LA 50020; Upper Group A LA 50021; Group I LA 50022; Lower Group F LA 50023; 
Tsankawi Group LA 50024. In late 1987 I was informed that these numbers were already in use 
in the LA system and that the official numbers would be Upper Group M LA 50972; Upper 
Group A LA 50973; Group I LA 50974; Lower Group F LA 50975; Tsankawi Group 
LA 50976. All handwritten records and photo boards etc. reflect the field LA numbers. 

Cavate group 

Hewett group, our subgroup, or other identifier. 

Number 

Assigned to individual rooms by this project unless Lister's or some other number is 
known (see Tsankawi discussion). 

Photo key 

This entry was not used in 1986 due to the lack of appropriate photos; it was intended to 
refer to a frontal location photo. 

Exposure/orientation 

Taken by placing Brunton in the center of the main cavate opening and reading the 
azimuth; significant shading, etc., of opening should be noted. 

Height above PGS of base of opening 

Vertical distance; intended to monitor how high one would have to climb or build to 
reach the opening. If the opening is buried give as negative value; -99 if not reasonably 
estimable. 



226 CAVATE STRUCTURES 

Fill Type: 

clear floor 

1 disintegrated tuff 

2 dung + tuff 

3 aeolian/alluvial 

4 high organic content 

5 tuff + dung + organic 

6 aeolian/alluvial + tuff + organic 

Approximate depth: Fill depth is frequently variable across a floor: attempt to estimate a mean. 
Use a probe in rooms with considerable fill. 

Estimate of completeness of room 

1 complete 

2 greater than 3/4 

3 greater than 1/2 

4 less than 1/2 

Evidence for construction 

none visible 

1 digging stick-like marks in ceiling 

2 shaping of opening 

3 digging stick + shaped opening 

4 shaped corners 

5 digging stick + shaped corners 

6 digging stick + shaped corners and opening 

7 floor-leveling fill (beneath plaster floor) 

Estimate of natural versus excavated space 

1 completely excavated 

2 at least half excavated 

3 at least half natural 

4 mostly natural 

Masonry 

absent 

1 single thickness large tuff blocks 

2 small tuff chunks with abundant mortar 

3 blocks present but not retaining visible coursing 

Cavate rooms in group 

Give the number of cavate rooms up-canyon and down-canyon in the whole group (not 
the sample group but, for example, all of Group A). This should include partial cavates which 
might not be fully recorded by this study. This observation relates to the apparent pattern that 
larger, "kiva" cavates seem to occur toward the ends of groups. 



APPENDIX 1 227 



Stability 



1 apparently stable 

2 deterioration evident, lesser threat 

3 deterioration evident, continuing, greater threat 

4 major problem, collapse or other major loss may be imminent 



Unusual tuff characteristics 
none 



1 softer than usual 

2 harder than usual 

3 unusual variation, stratification 

4 many large fibrous chunks 

5 highly porous 

6 soft red tuff 



Nonhuman use 

none visible 

1 ungulates 

2 pack rats 

3 bats 

4 insects 

5 combinations 

6 birds— raptors and vultures 

7 birds-other 

8 other rodents 

Inferred room function 

1 habitation 

2 storage 

3 "kiva" 

4 successive uses: both storage and habitation 

5 insufficient evidence for inference 

These functions are obviously normative and based on limited, perhaps unrealistic, 
attributes. Thus, habitation rooms are intermediate in size, smoked, and usually have vents, 
while "kivas" have more rock art and loom anchors, and are larger. Storage rooms usually are 
so called because they are not smoked, have less attention to plaster, and fewer features, and are 
often smaller. Refinement of these labels is highly desirable, and recorders should be mindful of 
ways of doing so. "Insufficient evidence" should be used sparingly. 

Room notes 

This is a space for verbal annotation of the base information code entries and about the 
room as a whole (for example, why is it a storage room and what is the imminent threat?). A 
word or two on the location relative to other rooms can be useful when sorting through a large 
pile of forms. Remember that there is no need to talk about features-each one gets its turn on 
stage individually in the next section of the form. 



228 CAVATE STRUCTURES 



Features 
Though subdivided here for the purpose of the form, every entry here is a feature with its 
own code, which is in turn described by other codes and measurements. Not every modifying 
code will be applicable to every feature. It is possible to take verbal notes for each feature; if 
you do so, mark the code line with a prominent N so that the person entering the data knows to 
look below for the note. 

Feature numbers 

We have two types of feature number plus the feature code: every feature receives its 
own number which is sequential within the cavate record, designated on the forms by "No."; this 
number is entered first, and when the form is finished the highest "No." should be the same as 
the total number of features in the room. The other type has to do with duplicate features in the 
same room (e.g., firepits 1 and 2); it is designated by #. There was some confusion as to wall 
numbering in 1986; each wall has a specific location code, leading to the logical (but wrong!) 
conclusion that there is right wall 1 and back wall 1 . The course followed instead is that there 
are several natural walls in the structure and they are numbered sequentially and located by 
chamber location codes, just as are other features. 



Feature Tvpe: 




1 


chamber 


2 


exterior door: doors opening onto the canyon; this will include doors that 




open into now missing masonry rooms. 


3 


floor 


4 


firepit (formal, lined) 


5 


floor burn (not a formal firepit) 


6 


floor ridge 


7 


subfloor pit 


8 


large floor-level niche 


9 


wall niche 


10 


slot 


11 


viga hole 


12 


smaller (latilla?) hole 


13 


loom anchor 


14 


possible upper loom support 


15 


smokeholes: often determinable by placement, angle, smoking 


16 


vents other than 15 


17 


natural (i.e., tuff) wall 


18 


masonry wall 


19 


interior door: doors opening into another room. 


20 


geometric petroglyph 


21 


zoomorphic petroglyph 


22 


geometric pictograph 


23 


zoomorphic pictograph 


24 


indeterminate petroglyph 


25 


indeterminate pictograph 


26 


exterior opening: the present openings of many cavates could not now be 




considered doors but need to be recorded 


27 


ceiling 


28 


indeterminate hole 



APPENDIX 1 229 





29 


incised dado: wide recessed band around base of chamber 




30 


beam seat: lacks pairing, opposites, other features of viga holes 




31 


groove 




32 


handprint (define medium in notes) 




33 


hand/toe hold 




34 


anthropomorphic pictograph 




35 


step 




36 


floor depression 




37 


wall depression 




38 


anthropomorphic petroglyph 




39 


chamber corner 




40 


metate rest 




41 


wall ledge: possibly for supporting a roof 




42 


passage 




43 


combination masonry and tuff wall 




44 


deflector 




45 


floor pit complex 




46 


posthole 




47 


vertical hole in ceiling (Panowski hole) 




48 


narrow wall incisions 




49 


cliff niche: large shallow rectangular niches not inside rooms 




50 


axe groove 


Doors 







Defined as holes large enough to go through, leading either to the outside or to another 
chamber; exterior openings are also recorded here. Another chamber does not include the large 
wall cists with large openings which will be considered features. Doors connecting rooms should 
be recorded only once and merely noted in the other room. 

location : code taken from chamber codes and measured by azimuth from 

center of the chamber to the center of the door 

shape : see list of plane shapes 

w.h : width is always horizontal here, height vertical 

wall thickness : measured at mid-door 

to room : give cavate number 

Chamber 

This records the attributes of the main room by wall. The numbers on the following 
chamber features are used to give locations of features, such as niches and doors. 1 & 2 are 
presently not used. 

3 Complete chamber : this code is used when the whole chamber can be described by a 

single form 

4 Chamber base : in recognition that some chambers are best measured as the 

combination of two geometric solids, base and top allow splitting 
of the interior space 

5 Chamber top : the base of this imaginary solid should, of course, match the top 

of the base 

6 Floor : refer to plane shapes, give dimensions of extant floor 

7 Exterior : the side closest to the canyon, in most cases with a door 

communicating with the canyon 



230 CAVATE STRUCTURES 



8 Rig ht: the side to the right of the exterior door 

9 Back : opposite the exterior wall 

10 Left : left of the exterior 

Cavate rooms do not have four walls in the same sense that masonry rooms usually do; 
some arbitrary separation of walls is required, but there is usually some basis for it. Some small, 
spherical rooms seem to be best treated as having right and left walls only. If there is no 
entrance from the canyon (i.e., a room opens only into another cavate), we will treat the entrance 
as "exterior." This may happen, but I think it will be very rare. 

11 Ceiling : areas need not be recorded here because measurements from 

wall tops have already been taken, but smoking, features, plaster, 
and damage are to be observed 

12 Exterior-right corner 

13 Right-back corner 

14 Back-left corner 

15 Left-exterior corner 

16 Within another feature 

Shape 

See list of geometric shapes. 
Measurements 

a,b,c, and d are defined for each shape as necessary 

Geometric fraction (Geo f) 

Refers to fractional shapes, for example 0.3 of a hemisphere or 0.5 of a truncated cone. 
Obviously these are estimates; they control for measurement of irregular shapes and estimate the 
area actually present (see also discussion in text). 

Plaster 

Coats : give total number of coats evident 

Smoked : give the number of coats that show smoking; it is possible, clearly, to 

have coats of plaster and still have smoking (recorded as coats, 1 

smoked) 
Color : 1 tan 

2 yellowish 

3 reddish 

4 several 

5 brownish 

6 black 

7 white or grayish 

Height: a very consistent feature is for the plaster only to extend partway up the 

side of a room; if possible record the intentional stopping height; for 
computer purposes, note that plaster height and height above floor for 
features such as niches (see below) are distinct variables. 



APPENDIX 1 231 



Damage 

This is a two-column code; it consists of natural damage and historic human damage. 
Human : absent 

1 minor graffiti 

2 major graffiti 

3 graffiti plus other modifications 

4 obvious wear 

5 serious wear plus other effects (tourist blasting) 

6 minor vandalism 

7 heinous vandalism 

Natural : absent 

1 moderate wind/water damage 

2 severe wind/water damage 

3 moderate cliff deterioration 

4 severe cliff deterioration 

5 cliff deterioration and wind water-moderate 

6 cliff deterioration and wind water-severe 

Features 

Numbering as above, and selection of type from the list. 

Location 

To the left of the slash give the chamber feature code; to the right give the TN azimuth 
from the center of the room. Beginning July 14, the practice was adopted of noting measurement 
of distance from Brunton to measured feature for at least three of the chamber aspects (e.g., 
chamber corner or wall midpoint) in order to have angle and distance, which will allow better 
reconstruction of where the azimuths were taken from and the meaning of the azimuths for the 
other features, without having to take measurements for every feature. Two types of features 
present special problems for location recording: those in the center of the room and linear floor 
features (such as the ridges occasionally seen on floors). The center of the floor will be 
designated 999°; azimuths for linear features on walls and floors have been left blank. 
Shapes and measurements from list of shapes; measurement d is for height, thickness, depth as 
applicable, unless a specific order is required by the shape being measured. 

Height above floor 

Measured to the base of wall and ceiling features from the floor; give negative value if 
the base is below floor level. In rooms containing enough fill that the floor could not be 
reached, measurements were taken to the top of the fill. These heights can be corrected by the 
computer by adding in the fill depth, or screened out in cases with a great deal of fill. 

Feature Fraction (Pea F) 

This is again a judgment, though a more subjective one than geometric fraction. Feature 
fraction is an estimate of the portion of the prehistoric feature that we have been able to measure 
(whereas geometric fraction is an estimate of the portion of the geometric solid chosen that is 
present). The value can be 1.0 for whole features, or less than or greater than 1.0, depending on 
the kind of deterioration present. This gives added detail to the overall completeness estimate for 
individual portions of the room. It is helpful to think of this estimate as a confidence band- 
features with FF's greatly deviant from 1.0 can be omitted from calculations. 



232 CAVATE STRUCTURES 



Notes 



Keyed to sequential feature numbers; it is useful to note associations with other features, 
especially viga hole pairings. In recording features it is possible to succumb to a sort of tunnel 
vision in which the feature is all the recorder sees, or at least records. Thus, while we want to 
know about each loom anchor, what is more important to the big picture is how wide the loom 
was; put that important, if infrequently encountered, information in the notes! (this is a true story 
based on real recording). 







List of Shapes 




Plane 












a 


b 


c 


d 


1 rectangular 


horizontal 


vertical 




* 


2 bowed rectangle 


horizontal 


vertical 




* 


3 oval 


horizontal 


vertical 




* 


4 trapezoid 


base 


height 


top 


* 


5 round 


diameter 






* 


6 triangular 


base 


height 




* 


7 natural oval 


horizontal 


vertical 




* 


8 linear 


width 


length 




* 


9 irregular 


width 


length 







*Measurement d will designate height, thickness, or depth for features or wall thickness at mid- 
door for doors (a plane shape with a depth is obviously a solid if one of those works better). 



Solid 





a 


b 


c 


10 rectangular 


width 


length 


height 


11 cylindrical 


diameter 


height 




12 hemispherical 


diameter 


radiusfh) 




13 truncated cone 


base d 


height 


top diameter 


14 truncated pyramid 


base w 


base t 


top I 


15 cone 


diameter 


height 




16 irregular solid 








17 sphere 


diameter 






18 pyramid (4-sided) 


base w 


base I 


height 



height 



Volumes for component solids will be calculated and summed by the computer. 



APPENDIX 1 233 



Added Attribute States by Date of Addition 

Note: This information is presented with the date of addition because it conditions when certain 
codes could not have been utilized. 

7/8/86 

Fill type 5 tuff-dung-organic mixture; it should be noted that the dung-tuff order does not imply 

relative quantity 

Evidence for construction 

4 shaped comers 

5 digging stick marks plus shaped corners 

6 digging stick marks, shaping of opening, and shaped corners 

Tuff characteristics 

4 high clastic content-leads to greater friability (seems very common at Group M) 

Feature types 

26 exterior opening (not a door) 

27 ceiling (redundant with location, but needed) 

28 indeterminate function hole 

29 incised dado-wide groove parallel to floor (M2) 

30 beam seat-hole large enough for a beam or pole but somehow apparently not a viga 
(position, pairing etc.) 

31 groove 

Plaster color 

5 brownish 

6 black 

Solid shapes 

14 truncated 4-sided pyramid (a=base w, b=base t, c=top I, d=ht) 

15 cone (a=diam, b=ht) 

7/9 
Function 

5 insufficient evidence to infer 

Features 

32 handprint rock art 

33 hand-or-toe hold 

34 anthropomorphic petroglyph 

35 step 

Chamber location 

12 corner between exterior and right walls 

13 " " right and back walls 

14 " " back and left walls 

15 " " left and exterior walls 



234 CAVATE STRUCTURES 



7/9 continued 
Solid shapes 

16 irregular solid 

7/10 
Feature type 

36 floor depression 

37 wall depression 

38 anthropomorphic petroglyph 

Chamber location 

16 within a feature 

7/13 
Feature type 

39 chamber corner (especially for locations) 

7/14 
Feature type 

40 metate rest 

7/15 
Feature type 

41 wall ledge (as for roof support) 

7/16 
Feature type 

42 passageway 

7/17 

Chamber location 

17 overhang-as on a projection past the front of a cavate 

Solid shape 

17 sphere 

7/21 
Feature type 

43 combined masonry and natural wall 

7/22 
Masonry 

3 few blocks with no visible coursing remaining 

Plane shape 

9 irregular (give length and width anyway) 

7/24 
Feature type 

44 deflector (just inside door, adjacent to probable hearth) 



APPENDIX 1 235 



7/31 

Tuff characteristics 

5 highly porous, vesicular 

6 very soft, coarse, and red 

Feature type 

45 floor pit complex 

46 posthole 

47 vertical ceiling hole (Panowski hole) 

48 narrow wall incisions 

Plaster color 

7 white to gray 

8/1 
Feature type 

49 cliff niche (large rectangular niches as at Tsankawi) 

50 axe-grinding groove 

4/2/87 
Feature type 

51 Brunton station location point not otherwise coded as a feature 



236 CAVATE STRUCTURES 



Materials Collected in 1986 



FS 1 
FS2 
FS3 
FS4 



Cavate 


Date 


M-37 


7/11 


M-61 


7/15 


M-65 


7/16 


TS-43 


8/12 



Item 
Possible human coprolite 

Possible human coprolite 

Fused carbonized corn with kernels attached 

Human infant bones: right (?) femur, fibula, 
mandible fragment, rib, temporal (right?) 



FS5 



A-57 



8/14 



Corn cob and possible squash rinds 



Appendix 2: Data Sets, Volume Calculation, Output Listing, and Photographic Data 
Documentation of Data Transfer from UNMto NPS 



The following data set members were transferred to the NPS system. The SSD 
extensions were generated by SAS: 



CAVBAS.SSD 

CFEAT.SSD 
CAVEAT.SSD 

NCBS.SSD 
NCFEA.SSD 

NONCV.SSD 
CHAMBER.SSD 



This data set contains base information on each structure recorded as a 
cavate. It includes location information, fill and condition information, 
date spans, and notes (approximately 144 kilobytes [K]). 

The full data set for cavate features, including measurements, location, 
shape, type, notes (around 2.1 megabytes [2100 K or 2 Mb]). 

An analytical data set for cavate features in which base information is 
included for each feature, and from which verbal notes have been 
removed; all variable values are numeric (about 1 Mb). 

Base information for noncavate features (about 146 K). 

Noncavate features; as discussed elsewhere, some of these features are 
recorded in groups with size ranges (404 K). 

An analytical data set for noncavate features that includes only features 
recorded singly and does not include verbal notes; all variable values are 
numeric (131 K). 

A generated data set in which each line represents one chamber; the 
variables are locational and give numbers for various grouped feature 
counts (about 76 K). 



Two tapes are being kept by the National Park Service Santa Fe office. Both are in SAS format. 
One contains the above data sets in AOS/VS (DG) format and reflects corrections and additions 
to the data sets after transfer to the NPS system. This tape also has some SAS programs used in 
the analysis on it. The second tape remains in OS (IBM) format and contains the data sets as of 
January 1988. 

The following data sets are in DOS Personal Computer formats on Vh" disks at the National 
Park Service: 

SPSS-PC +: 

Data: 

ALLCAVAT.SYS: 2845 cases from cavates; 47 variables, some labelled 

ALLFEAT.SYS: 3396 cases individually recorded features; 38 variables 

NONCAVAT.SYS: 551 cases from noncavates; 31 variables, some labelled 

CHAMBER. SYS: 130 cases combined and corrected chambers only; 47 variables 



237 



238 CAVATE STRUCTURES 



SPSS Formats and value labels: 

CAVFMT 

NONCVFMT 



ASCII: 

These files are ASCII versions of the above. They are in fixed format, with "." as a missing 

value marker. 

ALLCAVAT.DAT 

ALLFEAT.DAT 

NONCAVAT.DAT 

CHAMBER.DAT 



SASPC 

CAVEAT.SSD: cavate features with base information (similar to ALLCAVAT.SYS) 

NONCV.SSD: noncavate features similar to NONCAVAT.SYS 



APPENDIX 2 239 



SAS Program for Calculation of Volumes and Areas 

Note: Instructions are upper-cased and annotations are lower-cased following the line they 
describe. 

IF (SHAPE=11 OR SHAPE =12 OR SHAPE =15) AND MB = THEN MB = MD; 
Cylinder and hemisphere depths were variously recorded as MB or MD; this assures that all 
depths will be found in MB. 

IF SHAPE=10 AND MC=0 THEN MC=MD; 
Again standardizes the location of depth measures in MC. 

IF (SHAPE=14 AND MA>MB) THEN ME=((MB/MA)*MC) 

IF (SHAPE=14 AND MA<MB) THEN ME = ((MA/MB) *MC) 
Generate the width of the upper base of a truncated pyramid assuming that the bases are 
proportional. The conditional should not be necessary if the longest dimension of the lower base 
is entered first (as intended); this did not always happen, however, and the conditional assumes 
that the longest dimension of the top base is measured and will be proportional to that of the 
lower base. 

IF SHAPE=18 THEN VOL=((MA*MB*MC)/3)*GF 
Volume of a pyramid. 

IF SHAPE=17 THEN VOL=((3.142*(MA**3))/6)*GF; 
Volume of a sphere times the GF. 

IF SHAPE=15 THEN VOL=(1.047*((MA/2)**2)*MB)*GF; 
Volume of a cone times the GF. 

IFSHAPE=14THEN 
VOL = (MD/3((MC*ME) + (MA*MB) + SQRT(MA*MB*MC*ME)))*GF; 
Volume of a truncated pyramid times the GF. ME is the generated width of the upper base (as 
above), and the square root of the areas of the bases is the mean proportional. 

IF SHAPE =13 THEN 
VOL = 1 .047*MB*(((MA/2)**2) + ((MC/2)**2) + ((MA/2)*(MC/2)))*GF; 
Volume of a truncated cone times the GF. 

IF SHAPE=12 THEN VOL=1.047*MB**2*(3*(MA/2)-MB); 
Volume of a hemisphere. 

IF SHAPE=11 THEN VOL=((3.142*(MA/2)**2)*MB)*GF; 
Volume of a cylinder times the GF. 

IF SHAPE= 10 THEN VOL=(MA*MB*MC)*GF; 
Volume of a rectilinear solid times the GF. 

IF SHAPE =6 THEN AREA=(.5*(MA*MB))*GF; 
Area of a triangle times the GF. 

IF SHAPE=5 THEN AREA=(3.142*((MA/2)**2))*GF; 
Area of a circle times the GF. 

IF SHAPE=4 THEN AREA = (.5*MB*(MA + MC))*GF; 
Area of a trapezoid times the GF. 

IF (SHAPE =3 OR SHAPE =7) THEN AREA=((((MA+MB)/4)**2)*3.142)*GF; 
Circular approximation of oval areas by averaging the axes, halving the value, squaring and 
multiplying by pi, all times the GF. 

IF (SHAPE=2 OR SHAPE=1) THEN AREA=(MA*MB)*GF; 
Area of a rectangle times the GF. 

IF SHAPE < 8 AND MD>0 THEN VOL=AREA*MD; 
Calculates the volume for plane figures with depth measurements recorded. 

This procedure was carried out for all cavate features and all individually measured noncavate 
features. 



240 CAVATE STRUCTURES 

Retained Cavate Computer Outputs 

Note: These are on file at the National Park Service Southwest Regional Office, Santa Fe. 

I. Base Data 

A. Early Oracle output, uncorrected (1986) 

1 . Cavbase variables 

2. Noncavbase variables 

3. Cavate levels 

4. Noncavate levels 

5. Cavate exposures, fill depths 

6. Listing of cavate base information 

7. Tabulations of cavate variables 

8. Tabulations of noncavate variables 

B. Listings following successful conversion to UNM/IBM format 

1 . Cavate base data 2/ 19/87 

2. Noncavate base data 2/19/87 

C. SAS formatting 

1. Cavate formatting 3/9/87 

2. Creation of notes variable, trial list 3/13/87 

3. Catalog base data set 3/15 

4. Format noncavate base data 3/12/87 

D. Listings-Correction copies 

1. Listing of base data with corrections marked; notes A,F,I 3/14 

2. Base data notes for Group M and Tsankawi 3/15/87 

3. Noncavate base data and notes, corrections marked 3/16/87 

4. Addition of UTMs 

5. Addition of new LA number, renames field LA 3/4/88 

6. Line corrections 3/14/88 

7. Listing of cavate base data for Groups F and I 3/16/88 

8. Photo listing for cavate base data 5/3/87 

E. Tabulations 

1 . Cross-tabulations of cavate base variables 4/1 1/87 

2. Cavate HPGS by group and level 5/3/87 

3. Noncavate base data cross-tabulations 4/11/87 

II. Feature Data 

A. Oracle outputs, before data checking 

1 . Formats for CAVFEAT, NONCAVFEAT 9/86 

2. Frequencies, type shape, count, measurements, with and without nulls 

3. (4 outputs) 

4. Listing of cavate features, uncorrected 9/86 

B. Listings following conversion to UNM/IBM format 

1. Cavate features 2/20/87 

2. Noncavate features 2/20/87 

3. Generation to disk 

C. SAS formatting 

1. Cavate formats 3/11/87 

2. Noncavate formats 3/11/87 



APPENDIX 2 241 



D. SAS listings, correction copies 

1 . Cavate feature data 3/17/87 

2. Cavate feature notes 3/19/87 

3. Noncavate feature data 3/28/87 

4. Noncavate feature notes 3/28/87 

5. Listing of features with shapes 13 and 14 by BJM to check/correct 
measurement order 

6. Feature photo generation and listing by photo 5/1/87 

7. Feature photos sorted by feature type 5/6/87 

E. Feature tabulations 

1. Cavate and noncavate features by shape by group, by shape by part, by shape 
by cavate type, all features 4/19/87 

2. Cavate features, type by group 5/9/87 

3. Comparison of CFEAT+NCFEA with CAVEAT +NONCV data sets, feature 
type by group 3/3/88 

4. Counts of all noncavate features 2/24-6/88 

5. Cavate features group by feature type PC-SAS 3/3/89 

6. Noncavate features group by feature type 3/4/89 

7. Occurrence of viga and latilla holes by record no. 2/26/88 + other notes 

8. Niches-type co-occurrence, height above floor 2/24/88 

9. Milling feature co-occurrence 3/3 and 3/13/89 

F. Feature metrics 

1. Generation of volumes and areas, fill groups-test 4/18/87 

2. Volume and area means by type and shape 4/19/87 

3. Feature metric replacements, refined shape breakdown, counts by shape 5/1/87 

4. Door and niche dimensions 5/8/87 

5. Niche, beam seat, indeterminate hole dimensions 6/19/87 

6. Indeterminate hole dimensions and breakdowns 2/88 

7. Mean dimensions for conical and cylindrical features, wall and floor areas 
5/7/87 

8. Firepit and floor pit dimensions; plaster height to HAFL ratio by feature type 
6/19/87 

9. Fire pit and floor pit, floor depression dimension breakdowns 2/88 

G. Feature condition 

1 . Feature type by natural and human damage 2/26/88 

2. Wall damage by group and part; hearth listing 5/5/87 

3. Smoking on walls and ceilings by plaster, function, group 3/10/88 

4. Moderately and unweathered walls: plaster coats and heights, and plaster color 
by part, group, function 3/25/88 

5. Plaster color by group: all walls, back walls 6/9/89 

6. Floor plaster coats by group and function; wall plaster coats by features, 
smoked coats 1988 

H. Multivariate analysis of features 

1. NCSS Cluster analysis of indeterminate and latilla holes 6/18/87 

2. Cluster analysis attempt for several hole types 7/19/87 

3. Cluster & discriminant analysis: several hole types 3/4/88 



242 CAVATE STRUCTURES 



III. Chamber data 

A. Listings and attributes 

1 . Creation of chamber data set, some frequencies 6/10/87 

2. Listing of chamber data set, frequencies of attributes by group and function 
7/19/87 

3. Generation of location index, tabulations 2/26/88 

4. Repairs to data set 3/14/88 

5. Listing of corrected and updated chamber data set 3/23/88 

B. Tabulations and analyses 

1 . Chamber metrics by group, function, shape (from feature, rather than chamber 
data set) 6/19/87 

2. Shape by niche 2/26/88 

3. Means, breakdowns, group comparisons, t-tests for chamber volumes 2-3/88 

4. Means for chamber volumes by group combining habitation and "kiva" classes 
3/5/89, 3/9/89 

5. Tables of feature number by plaster coats, chamber data 3/14/88 

6. Co-occurrence of feature groups 3/21/88 

7. Feature occurrence correlations, all and selected chambers 3/23/88 

8. Location index cross-tabulations 4/4/89 

C. Multivariate analyses 

1. Discriminant analysis by function and group, first attempt 7/19/87 

2. Discriminant analysis by function and group, different variables 3/25/88 

3. Cluster analysis, Ward's method 3/3/88 

4. Feature cooccurrence and cluster analysis 3/14/88 

5. Fastclus with 6 and 10 clusters, different variables 3/16/88 

6. Chamber clusters with listings 3/18/88 

7. Fastclus analyses, listings different clusters and iterations 3/21/88 

IV. Data management 

A. Transfer from NPS format to IBM format 

1. First attempts 2/5/87 

2. ASCII tape with incorrect block sizes 2/18/87 

3. UNIX tape index showing incorrect block sizes from NPS 2/18 

4. Regeneration of UNIX data (M. Prine) 2/19/87 

5. Generation of data from tape to UNM disk 2/19/87 

B. Tape index for UNM formats 

1 . Tape on file at NPS in IBM format (CAVUNX) 2/20/88 

C. Transfer SAS formats and data to NPS system 

1. Six data sets to tape 12/16/87 

2. Seventh set to tape 1/14/87 

3. XCOPY for export data set 2/1 1/88 

V. Listings of corrected files 

A. Contents and listing of CAVBASE.SSD-cavate base data 

B. Contents and listing of NCBASE.SSD-noncavate base data 

C. Contents and listing of CFEAT.SSD-cavate feature data 

D. Contents and listing of NCFEA.SSD-noncavate feature data 

E. Contents and listing of CAVEAT. SSD~cavate feature analysis set 



APPENDIX 2 243 



1 . Frijoles data 

2. Tsankawi 

F. Contents and listing of NONCV.SSD-noncavate feature analysis (single features 
only) 

G. Contents and listing of CHAMBER. SSD--chamber analysis set 

VI. Various SAS GRAPH programs for generating figures 



244 CAVATE STRUCTURES 



Photographic and Videotape Data 
Photo Rolls, Formats, and Numbering Conventions 



4x5 
These were numbered only by group and exposure number, and were so entered. Photo 
A-2, for example, corresponds to the number written on the negative and the photo record. 

Rock art (35 mm) 
The Crowders assigned roll numbers for this series; they have had the prefix CRA 
(Cavate Rock Art) attached. CRA10:12,13 indicates black-and-white rock art roll 10, exposures 
12 and 13. 

Color slides (35 mm) 
There is only one roll of slides; it has been labeled CC (Cavate Color); CC1:5 is Cavate 
Color Roll 1, exposure 5. 

NPS rolls (35 mm) 
Four rolls were used during the recording session, two were shot before the season 
started, and three were shot in April 1987. All these rolls were given the prefix CAV; the four 
used during the season were CAV1-CAV4, while the two taken before were CAV8 (taken by 
Toll in April) and CAV9 (taken by A. Ireland, also in April 1986). CAV9 is all distant shots of 
whole groups from across the canyon, and was not entered on any data lines. CAV8 includes a 
number of shots from groups in which we did not record. Also included with this group are 
three rolls (CAV5-CAV7) taken in April, 1987 by B. Crowder and W. Toll to prepare frontal 
drawings, to rephotograph a few more cavates shown in Lister's photos, to fill in some gaps, and 
to reproduce Lister's 1939 photos of cavates in which we worked in 1986. Only a few of the 
April 1987 (CAV5-CAV7) shots were entered on data lines. 



Video Tapes 

Videotapes were taken for all cavate areas recorded. These provide color shots of all 
walls and features as of summer 1986, along with verbal commentary, except for a few tapes on 
which there were some audio problems (these problems were partially corrected by voice-over 
after recording). 

The tapes are in VHS format and are kept at the Southwest Region offices of the National 
Park Service, Santa Fe. 



Appendix 3: Base Information, Threatened Cavates, and Room Stability 
Computer Output of Most Base Information for All 1986 Cavate Forms 



Variable abbreviations are as follows: 



RC# 


Record number 


GP 


Cavate group 


cv# 


Cavate number 


ELEV 


Elevation 


EXP 


Exposure in degrees true north 


PGS 


Height above present ground surface 


FT 


Fill type 


FILDPT 


Fill depth (cm) 


COM 


Completeness 


CON 


Evidence for construction 


NES 


Natural vs. excavated space 


M 


Masonry 


UP 


Rooms up-canyon 


DOWN 


Rooms down-canyon 


STAB 


Stability 


TUFF 


Tuff type 


NHUSE 


Nonhuman use 


FUNC 


Function assignment 



On this listing zeros show as blanks 

Note: Variable values and definitions can be found in Appendix 1. 



245 



246 CAVATE STRUCTURES 



U.3ZU I — — CM — Ift — — — W — — — CM — — (0 — — CM — CMCM — — — — — — CMCM — — — — — — — — CMinCMCMCMCM — CM — IrtCM — 

I 

ZI3MU i * vios * « « in in in in « « cm co co «- oo — in»<*invwN^-in * t -<■<■* 

i 

t-3u.ii. i n co co *» co co co co » co co 9 * <■ 

i 

to »- « » i <m«m — ^cocmcocm^^cocococo- vcm — — nN(M^nN<-(ONNNNNNn«NN«nNNnNNnNnnN 
i 

z ■ oi«inoaiinnoieioioifiin-<»noiooa)<tcooiMO'M^ift-oMO)U)-cnB- ioii)o>-OMOnnci 

3ti «m(m(mm — — — ooooanatBocthhhsitiBiBitttnnnnNr — — cocococococmcmcmcm — 

ol ___________ - 

i 

a i — ^oio^ioooo^ — cocouj^^r>>«jr»ocouoo>- tooior^cMcooof^cMcowo — cm v <o t^ on — nnid 

oi •-•-•--MNNnnnnM«4iAiAiotsiO(ONViniAtAioiO(aifiMoa)iso — — — — 

i 

i 



Z lil l/> I — — — CM CM CM CM — CM — CM CM CO ^ — CM — — CM CM CO — — — — — — CM — — CM — — CM — CM — — MM — — — CM — CM — — NN 

I 

uoz i '<«- in — * r^ ^ * in mo oitmig co in in in co <o — — — * » «o in — <o <o — — co — * in in in—. 

i 

UOS I * CO — CM « CM ^ — * — CO «* CO CM — CM — — — — — CMCMCM — — CM— — CM — — — CM — — — COCOCMCMCMCMCM CM CO CO CM 
I 

_i(-io — in — ooocmok cm — o — o — cm — cm (o-Tointoo moo — o «co <o a<rincooinaocMO- 
-•aiOOOOOOOOOO OOOOCMOOOO ooooo — o o - — o — oo O O O ^ O O CO — — o — 



u. >- i — — — co co — coco co — id co id co — toco— — to — co od in co cm minco— in co * co — coco co 

w i i cd o in r- ooooino — ininininincMocMincoininoa>cooininoininoo)oo ooooo ooin 
Oio a>t>-oo) co — oino><o<fio>*r^o«)cocot>-r^inincocMa> — oocm — co«M^r>-ooo> a> ^ o ^ cm ooin 
a. 

I— CM — — — in — CO — mCM — CM — CM— CM — — — I — — II— CM (N CM— CM— — 

I 

a. i in — fooMnaiaininooNiniooN^iDiDNinnoaio- aootf>oinin<ocoovioou>a>ininoooino 
x i r- <r in f- id co — — i^cMr^coinco^' — oinv — cocmcocm — — viD^nnaoo- oi co cm eo — r» o> •- o — ^ ^ co *• w 

UJ I — — — — CMCMCMCM — CM — — CMCMCMCMCM — — CMfMCM — CM — CMCM — CMCMCM — CM — CM — CM CM CM — — — CMCMCMCMCMCMCMCM 

I 

> i oooooooooooooooooooooooooooooooooooooooooooooooooo 
uj i — — cmcmcm — — — co — — — cm — — co — — — — — — — — — oooouJr»N(C(0(0<oiOf«'iot»^^^ , v»nioinininininin 

_l I CMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMNCMCMCMCMCM — — — — — — —— — — — — — — — CMCNCMCMCMCMCMCMCMCM 

oj i <^(p(0(0(omco<n(0<0(OiO(OtototDtD<DtoiDtO(om<o<D'i3(OiotD(n(i3toto(om(n(n<i)<ocDtomco<Dmu)co<n<Dco 

i 

* i — cm^-o — nma)ONBOiON»oiNOKON(ON- cocM^oicMinioco^ — inr^ooin<oo»cMco»nr»c»ocMco^r» 

>l — — — — — CMCMCMCMCOCOCOCO*inWtO<0<0<Or-r- — — — CMCMCOCOCOCO^inW _____ 

U I 
I 

a i 

13 I <4<<<<<<<<<<<<<<<<<<<<<<<ILU.U.U.ILIII1.ILILILILILU.IIU.HMHHHHHHMH 

I 

I 

I 
_ I 

uj i i- i- azca a a. a. o i- t- t- t- a. o i- co a co no oo 00200000.0000 

OIU.U.S S a. -> a. a. a. a. ofu.2u.su. u. 3 3 a or jaatraijLCcau-crTQractctaccccceo; 

x 1 ami x oou o 00 uo I o I o in 01— xuioujuioujujoujoujujujujoujujiijuj 

u 1 i-i-o.i-ino.t-i-ijhiahaiaiaiiiiit-ii-iifflhiizi-iiaiffllllii-iiil 

oc 1 xxoxoxooxxaoxooxooaonnaaaxaxooujxaaoxaooaujooaoxoooo 

I (D1O<D(D(OlD(D(DttlDIDlDIO<O(OIO1O(OIO(O(fiCD(O(O(OU)U}IO(O(Dtf)IO(0(DI0ID<DIO(DID(OtDID(0<DID(D(OIOID 

I 00 CO CD 00 00 CD CD CD CD CD CD CD CD CD CD CD CD CO CO CD CD 00 00 00 00 00 00 00 CD CD CO CD CD CD CD OD 00 00 OD 00 00 OD ^^ QD OD CD CD ^^ CD 00 

I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 

I — -I -I -I -I -I -J -I -I JJJJJJJOUUI9I9I90UI3J -I _i -I -I -J -I -I _J_ _ -I -I -1-1 -J -J -J J J -I -I — -I -I 
l33_33333333333rJ!3r5r5D3D3333333333_D_3DD_3333=>333D3D_ 

UJ I -»->->->-»->->->->->-J->->->->-5<<<<<<<<<->->->->->->-»->->->->->-»->->->->->->->->->-J->-> 

I- I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 

< i <o r~ cm » r» — — — — — — CMCMCMCM — 9«««vu)<Tinvooooooooo>o>o>(jiO)0>o>cococotococococococo 

Q I — — CM— — CMCMCMCMCMCMCMCMCMCMCM — — — — — — — — — CMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMNCMN 

<t i r-.oD<om<oortin<floair-Taio(^o«TOCM(o<T«)(Nr-^-(-.cMui — a> f- «o ^ o — co <p r- o oo on — r> <r <o r- o co t 

o i ioioossa)a)ODa!(iiii)»ocno(DP)n<nf<iin<nii>»'r*ifliniOiji(OiflMniBiDa)*ooo--t--NNN 

a. i — — — cococococococococo — — — — — — — — — — — — — — cm — — — — — — — — — — 

i 



APPENDIX 3 247 



U.3ZOI co — — — cm win — — cm — cm — — cmco — inwin — in — — in — --mm — co — co — cm — — — cm — — in cm — cm — — 

i 

•-MM 

I3MUJ i io-tii) icwmuiin r»T* * v p» » v » ^ r» r» in » in in «• «■ ^ v 

i 

10 

\- 3 u. u. i co ************ <» * « — t * <r •*"» t ^ v <» ^ 

i 

i/> »- < co i — * «r — co co — *• cm co * -^•-Nn^n^nnfxnno'-NNn-NB'-nn^nnoonnNn^ 

i 

cm — — 

zi en m — — «»'-oiB^»oo«)ooioi"-NiOMO((ino)oiMfliflN«oUN<finonNOMOin*<N- 

ji .- — — — — nonooNNN'-'-NN'-'-'-'-'-ooooooooiJiooiuioioioiiniBOfflBaiin 

i 

a. i io 05 oi co n cm cm <o — comiooo^cjicniovcM — cM^Mfvoo — cocoinincoeftinencMcocMf^cooco^iniOflo 

o i — — — ts cm co co co t-^r-^t-i-f-NNNNNNNr)onnn(onnn<f<<tttinminiflifliri 

i o a> o 

I CO CM CO 



ZU1V) I •-•■ — — CO CM — CM N CM •"N-'-NNNN- CMCM — CM — — — — CM — CM — — — CMCM — CMCMCMrMCMCMrMCM — 

I 

CM — — 

uoz i itiAioiDin — ioism in — co — — * co * *■ co co co ^ io * ^ — * ^ 'f it'* 

i 

UOII <■'•■•■« l>l-NNNt — CM — — ^^CM^-^^CM**^* — — ^COCMCO — CMC0CM^C0***^CM*CM 

I 

n — — 

ji- i ninvrto eninooowioinocMcocM noN-ooN-oN'- o o o cm — oooo — — — co — — 

■-• 0. I O CM O O O WO — CMOOOOCMOOO OOOOOOOOOOOOOO— OOOOOOOOO OO 

U.QI CROO 

I O — CO 

I ... 

u-Ki in « co — — — co in in — co — — — — — — — — — co — — w — — — • -co 

i 

co in co 

i/i i o o o o o oenwcMOooooooooooaooiniDooinoooiocMoo* moooooooooo 
d i aio-'h NoimooooaiaiaiNOiNiflovnMiin^NOaqiAvn^ cooo- i^oicoo — co — co 
a i ooo 

I— — — CM O O CM I — — CM CM CM CM <T U"> 10 CM CM «» CO — CM CO CO CM — CM (N ^ — CM CM ^ in ^T CM 
I 

o. i t in in in in o in — — o an in — ^ » — — in ao in co o cm ^ m-eooois- in — oinoin cm 

x i in ^ ^ ^ <■ cm r» 10 co co CM^-cotcom cm — in in o o co r» coco — ^coooJcocMmio^iooQ v 

UJICMCMCMCMCMininCMCMCMCMCMCM (MCMCMCMCMCM CMCMCMCMCMCMCM CM CM CM CM CM «M CM — N N N N (M N N CM 

I CM U> CM 

CM — CM 

> i ooooo ooooooooooinoinooointnoooinooooaininoinininmoinominoino 

uii in in en in to kii/ii(inn-----oo'-nnoooO'-'-0'-0"00o-^-0'-o^oo-oo 

_J I NNNNNOOONNNNNNNNNNIMNNNNNIMNNNNNNNNNNNNNNNNNNNNNNNNN 

uj i (o<oiO(O(O(ninin(OioiO(OioioiciD<otD(O<O(O(0iD(OiD<oi0(Oioio<om(Dioio(0(Diocoio(D<£iom(O(ommtoco 

I CM CM CM 

10 

* i aioNios n»si- Nn<fionontifliDNao)a<- nvuimo- coTiniot^coo — cm^-cooooio — cmw 

> I — CM CM CM CM C0COC0 — — — — — — — — CMCMCMCMCMCMC0C0C0C0C0COC0COVVfV«<rVininirm 

O I O — CM 

I CO CO CO 

a. i 

O I ■-• mm mm mmmZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZ 

M M M 
I 

I 
I 

<r I u_ 

u i ninos to o, o> z a. a z t- t- hi- a a a 

a i a -> or a u.u.u.ora.o.-)a,u-u- a -> 32 )) a a. u_ u. a 

aciiDffliiJiuintDnuiisiocDcDfiOD co co XX xx co oo co id 

O I >» 'V "* >» V, ^ V ^ •>• ^ >» ^ ^ ^ v. V ^ v. >^ ••». v. v. ■v. -^ 

u 1 »-aazza.aaazi-)-(-i-zzi-ah-t-azzZA.zzo.i-Ka.ZZi-i-CLZZt-ai-t-ZZa.aa.CLa.ai- 

a 1 xuiDcoocoococaxxxxomxoxxocflocacDiDmcoxxocDoxxcocQcaxcoxxiDaicDiDCDaicficox 

I <O0tO(OlOtfU)lOU)lO<OlOlOll3(O<OlC(O(O(O(OlO(O(O(OlOlOlOlOl()(OtOlOlOlOlOlOlOlOlOlOlOlClOlOlO^OlCl0li3 

I 00 CD QD OD QD CD CD 03 CD CO CD 00 OO 03 00 CD 00 03 CO 00 00 OO 00 CO 03 GO CD OO 03 CO CO 00 03 OO GO 00 9) ID ffi 00 CD CO OO 00 CD 00 00 00 00 

i i i i i i i T i • i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i 

i j_i_i_i_ij_i_i_)_i_i_i-j_)-j_)-j_i_i_)_i_ij_)_i_i_i_i_i_i_i_i_)j_j_i_)_i_/_i^_i_i_i_i_j-i_i-j-J 

103033030333033033000033303033030030030303030300333 

uj I ~>->->-y-i-y-y->->-)~>->-»->-y-)->-i->->->-i->-y~i->~>-i->~)->->->-i~>~>->->-i-i~>~i-i-)->~>~i~i~>~> 
t- I I I I I i I I I I i I I I I I i i i i i i i i i i i i i i i i i i i i i i i i i i i i i I i I I I I 

< i nnn7««Y4tt«aoaooaoocomioiciDi»(iio>oioao)aia)0<aooo^^«««««*t«viniAin 

Q I CMCMCMCMCMIMCMCMCMCMMOOOOOOOOO — — — OOOO — — OOOO — — — — — — — — — — — — — — — — — 

% i cMuir» — CMoDin»t>oo — cMcoMfooc»onvin^cMin<ot»o3ai»^oco^inrMoajcoto^o — coo)CMin«or^ — cmo 

<_> 1 CMCMCMCOMCOCOMCOO* — — — (010(0— — — — CMCMCMCMCMCMCOCOCMCOCOCO^MrCOCO^^^-^inWin 

OK I — ______,-.-,()- 

I 



z 


a 




-> 


a 




co 


CO 




■N* 


>. 




a z 


K K 


H- H- 


a. -> 


* 2 


i * 


CO CO 


X I 


I X 



248 CAVATE STRUCTURES 



U.3ZUI— CM — — — CM CM .- .-.-m — — CM — — — — — — — CM — — CM — — CM — — CM — — - — r- — NNO-- — -CO — CM 

I 

l3Mui i r>- v « * — in < 10 in co v *• — — * in in * — — — — in — <dcm eo t^ oo * * 



t- 3 U- U. I 
I 



in in in in in m in m m in in id id io to id id id is <o <o to <o io to to (o to in mmm 
in 



l/l)-<ID I 4>-NN(M<- — CM NN(Mt-N»NNWIM^nr)NNNN>-nN(M<feMNrK nnnN--NBNNt-<-nN(M 

I 

M 

z i e»>e»r»*co*cocM 
5 i r-.r-r-r-r-f-acM 
q i — 

i 

a I O)0IN7IA«N« — 

3 i mmtotototo* — 

i 
i 

Si— co - 



ZIUI/l I N<-NN>-- — — CM — CMCM — — CMCMCMCM' 
I 



CM I 



CM 



CM CM i 



(N-(N (N — C4 



uoz i iflNiniflinn m — «o — co — — — in — «» — * in — co in — cm co cm — co co co in to — to w cm 
i 

OOSMV — — CM — — — CM CM — IM — (OMCMCON — — WV — — — M — — <•> — — CM — — — MCM — — — — CM CM CM — CO CM CM 
I 

j(- i on cm — r» mo in to cm cmocococmuicoo oinvininr»ooinooooina>iDoininrM--o- ooio 

-• a. i o o ooo o— o— o ocmcmooooo ocm — f» r» — ^moNininNO- ocm — ooooococo — 



u.f-i co cococo t <o coco cococoiotocococoto (OintoiotoiniocMin — in in in in — co co to to to to m nioion 
i 

CO 

t/i i co o o o o o oo r-ocooooooto intoooooo ocMoaomooooooto ootomoomom 
o i cm — to m in in oo nonooooos — ai^oooo ocm^co^coooodwoco oo — nnnnoki 

a i o 

icMCMcoininintoo— cm— co — — cm — — 

i 



a i in t r- t in in 
x i in co cm cm to CM 

ID I CM CM CM CM CM CM 



in cMf»cM — u> OO in to — ooaoinoooininococominoMCMininoinofoooico- cm m co m 
k cocooh-* — — oo>cM — N^mincMO — o — o>oocfto>r^o)oo>o> — cmooi — in^cMointa 

r-^^^.».-CMCMCMCM- — — — — — — — — — — — — — — — w- w- — ■-»-.- 



> i ooooooooo oooooooooooooooooooooooooooooooooooooooo 
uj i oooooococmo oo — — oooocnooooooooooooootscocoo- t^t^t^aoaaxDaasaxB 
_i i cMCMCMCMcMCMCMCMioocoioiOieiOiototoiniotoininininininininininin^^^^inininininininininininininin 

UJ I t0l0(Ol0(Ot0l0l£l0O<OCOlOCO(DtOtO<OlOtO(OlO(OtOU)(OlO<OtOtOlOtOtO(OtOlOl0<OtOtOtOtOtOtOtOtOtOtOtOtO 
I 10 

<0 

* i <0 r>» oo o> o — v in w r»a>o>o — covinton»coeno — CMcowintor-cocno- cm co ^ in o — Nra^mtomoio- co 
> i mininintoioioto — — — — cMCMCMCMCMCMCMCMCMcocococococococococo**«^**ininininininininin<o<oio 
U I <o 

I — 



I 3Z32Z3Z2I- 



t/> i/)ioi/ii/)vii/ivivii/)i/ivii/)i/iv)toi/ivii/iv)i/)vii'ii/ti/)i/ii/ivii/ii/ii/ii/ii/ivii/ii/ii/ii/iini/ii/) 



KI-t-l-»-»-KI-»-(-l-l-»-»-l-l 



I 
I 
I 

a i 
uj i 
a i 
<r i 
O I 
u I 

UJ I 

cc i 



-> I o- 3 ■ 

DIBIl 



com & a a. »- s t- Q- »- n- 

>>te a s: u. a a * -> u. 3 as 33 
z UJ (0 03 ■ uj ■ x id so I oa r IX 

UJ'vlL.'V^v.'*.^ v. x. ^. v. >.%» V.V. 

:luiaiiiiia.jiigit.iiaiiiLiiaii 
>->oco.->oK"^-^ao:->a:a->-»a-»-»a.->->Q.Q. 

IO3OflDiO3UJO3flDOBUJAuJfl3O9flDCDO3(DO3O3fl303O3 



Z O. Z 

-> a -7 

03 03 03 



os a 

Z Z 2 
03 03 CD 



U. X U. 
BID 
>, >. ^. 

Z a. z Z 

-> a. -> "j 

03 03 03 03 



Z a Z t- 
03 03 03 I 



I- a. a. 
* a. a 

X 03 03 



3 3 2 

III 

■v. v. x. 
Q. Z Z Z Z 

0. -> -> -> -» 

CD CD CD CD CD 



CO fp TE CO CD 00 CD CD CD CD CD CD QD CD CD CD OO Q3 CO CD CD CD ^JD ^D QD 00 00 CD CD CD ^0 35 OB 00 DP * T) CD CD CD 00 00 ^ fl CO OO O B ^^ 00 00 CD CO 

i i i i i i i i i • i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i 

JJJJJJJJJJJJJJJt3l90Jl3l}Ol30aeOt9Dl3t]l]l)OOl3l]lll}t}Ul3l}i:OOOt}0 
33333333333333333333303333333333333333333303333333 
->->->">-»->->->->->->->->->-><<«-><<«««<«<<«««•«<<<<«•««<<««<«•«<<<< 
I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 

in^r>-inininio«o — — — — — — — — — — — ^^k — — — — — — — — — cMCMCMcocMCMCMt^r^«r«-tototointoininin 

— OO — — — — — COCOCMCOCOCOPIOOOCOOOO — — — — — — — — — — — — o— — — oooooooooooo 



i ro co «■» t in to - coinior^cococMOD- cMco«oifli>inooo>o-cMco«rtn<o«ro>co»tina>ofxin<j , iococM — to o> *• co cm 
i inininininmiotooooo — — — cmcmcm — cococoa>a>o>oicncnci>cnoioo- — — — cMior»r»r»<otoioioinininin 

I CMCMCMNMNCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCOmCOCOCOCOCOCMCMCMCMCMCMCMCMCMCMCMCM 



APPENDIX 3 249 



U.OZO I n-nNWIONNNNNNNIONNNiniO'-WNNN 



zi3^u i * m 



i-3u.il i iflin louiniAifl in in ioioioiow 

i 

l/> H < CD I — CM CO — NONNOBNT-N- — ^ON- — —CM 



Z I 
2 I 

O I 

I 



a i 

3 i 



S I 



CM — CM CO CM CO CM - V — — CM — — — — — CN 



uoz i lonieiS'- — cm 
i 



to m co 



UOII "->-<-n"<nNN-B--N CM — — — — CM 



Jh I NOmOUOCOIOOIOOONOMONNOOinlO' 
•-• 0. I OCOO^OOCMCMOOO— —OfM — OOOOI^CMWO 

U. O I 



u.1- i nnioicn in in co -co <o <o — — nuiumu 



to i o in o o o o 

a i o - o o o - 

a. i 

I CM 

i 



r» a w o o o o in ooo^oo 

OltNOOOON <000-'»0 



ai o in m cm — * o a o> o o> coo> r- o <o co t «c a in in o 
x i ^ — cmcmow — <o — oo — tocom(ocM^Tt<oot« — 
ai I — — — cm cm — cm — — — — cm — — co — — — »- 



> I OOOOOOOOOOOOOOOOOOOOQOOO 

uj i aaoNo-oooooo^ifiaioMOoooooo 
_i i inininin<0<O(O<o<o<0<0(0inminininm<r*<*<r^'in 

I 

* i ^w<or» — f-eocn^inr^cn — coo — oocoaco^r^coo 

> i «o<o«o<ooooo — — — — cMCMcocococo^wcococor* 
u i inir>inir>inu?mir>mir>tnu)inmmif)inu)inui 



(/)(/)(/)(/)</)C/)</)lfll/)</)l/>(/></)</)(/)lft(/)(/)l/)(/>(/)t/></)</> 

t-t-t-t-t-t-t-t-y-t-t-t-t-t-t-t-t-t-t-t-t-t-t-t- 



a 
iu 

Q 

a 
o 
u 

u 
a 



a. 

u. a. 
to cd 



3 2 U. 
I ICO 



a. 

u. a 
m co 



ao.sszi-zsa.so.t-zzai-1-t-ixsszzz 

CDGDCDCDCOXCOCOCOCOCQXCOCOCOXXXCOCDCOCQCOfD 

I I I I I I I I I I I I I I I I I I I I I I I I I 
I OOOO-1-J-IOUOOOOOOOOOOOOOOO 
I 3333333333333333=)33=)3333 

Ul I <<<<->->-)<<<<<<<<<<<<<<<<< 

t- I I I I I I I I I I I I I I I I I I I I I I I I I 

< i ***co — — — — — — ****«o<or»r«-r»- cm — coco 

O I ooo — cocococoooooooooooo — — — o — 



e 

u 
a 

« 

M 

m 

T3 

L. 

o 



i o(MSN«a)N^Nniooi'-«ifl«iooNMOniri- 

I tt<NO'->-N«OO»**(fl»(0MB0)-OOCM 
I CMCMCM<OCMCMCMCMCMCMCMCMCMCMCMCMCMCMNN»inmm 



250 CAVATE STRUCTURES 



o 
Z 

s 

o 

3. 



1 
z 

E 
I 

u 



! 

o 
«> 

l 

a 




© 



CM 



0\ 
CO 



c 
o 



r 



i 

8 B* 
.s — 

"■* 8 

5 .S 

if 

cd J3 



T3 






g 
ff 






2 3 

Q- o 






§■8 * 

> > .g 



4> 
2 ^ 

* eg 



£ 



O Q 

•8-1 



2? 2 « 

18s 

• O £ 

1,11 

It-S 

o 



68 

© 

— CM 
CM 






a 2 "8 
S S J 

i-s §• 
I s3 § 

OS o O 



o 

CM 



tu ~ 



•5 3 



2 -a w •- 



P 8 "3 



CM 



APPENDIX 3 251 



o S j2 

» O g * 



.9g| 




- £-5 K * 



* <* £ 



cs 












s s s 



s s 






252 CAVATE STRUCTURES 



o 
Z 

E 
o 



M 



E 
o 
o 



03 
> 

U 



a. 

2 
a 



at 
% 

u 



i> s i 



S 



op 

1 g 8 < 

"C ■" =3 J= © 
P a i 8 

<fi '3 "8 £ 3 

JBiJf? 

-, «5 C » 

2P g «o S *> 

5 u J2 "3 
5 o. >>m >. 

3 8 -a -° 8 

w O 4> 

v «-• Z, 

00 I- u_ 

o 5 c 

« «- o 

bOJO 

5 2 o c -^ 

2 *3 M 

"> c o u- c 

U O O '" 

3 = S 5 M 



65 




5 ~ O « <„ 

u y jo 5 8 

£ 2 « S J = 

H o 35 -S 15 C 






CO 

H 



«n 



11 



3JJ 



00 



APPENDIX 3 253 



Breakdown o/Cavate Room Stability by Cavate Group 



Record Cavate Amount Excavated Human Natural 

Observation Number Number Stability Tuff Type vs. Amount Natural Damage Damage 











GROUP A 








Apparently stable 














1 


106 


4 


1 





1 





3 


2 


100 


34 


1 


3 


1 


1 





3 


334 


50 


1 


3 


1 


9 


9 


4 


340 


57 


1 


4 


2 





1 


Lesser threat 
















5 


67 


1 


2 


3 


1 


9 


9 


6 


68 


2 


2 





1 


9 


9 


7 


80 


13 


2 





2 


9 


9 


8 


85 


18 


2 


4 


2 


3 


5 


9 


330 


47 


2 


3 


1 


2 


3 


10 


345 


63 


2 





1 





5 


11 


354 


66 


2 





1 


9 


9 


12 


346 


67 


2 


4 


1 


2 


5 


Greater threat 
















13 


82 


15 


3 


3 


2 


2 


6 


14 


89 


23 


3 


3 


2 





6 


15 


104 


30 


3 





3 


9 


9 


16 


99 


32 


3 


3 


4 


1 


6 


17 


342 


60 


3 





2 


9 


9 


18 


343 


62 


3 





3 





6 


19 


347 


73 


3 





1 


9 


9 


Major problem 
















20 


75 


10 


4 





2 


9 


9 


21 


86 


20 


4 


4 


2 


9 


9 


22 


90 


22 


4 


3 


1 


9 


9 


23 


87 


39 


4 


4 


2 


9 


9 


24 


352 


71 


4 





1 


9 


9 



254 CAVATE STRUCTURES 



Record Cavate Amount Excavated Human Natural 

Observation Number Number Stability Tuff Type vs. Amount Natural Damage Damage 



GROUP F 



Apparently stable 




25 


147 


4 


Lesser threat 






26 


144 


2 


27 


155 


12 


28 


161 


15 


29 


159 


16 


30 


167 


23 


31 


166 


27 


32 


174 


31 


33 


183 


38 


34 


186 


45 


Greater threat 






35 


152 


9 


36 


180 


35 


37 


197 


56 


38 


200 


59 


Major problem 






39 


181 


37 



2 
2 
2 
2 
2 
2 
2 
2 
2 

3 
3 
3 
3 




















9 


9 
9 



9 

9 
1 
9 




9 

5 
5 
9 
9 

1 
5 
5 
9 

9 
6 
9 
6 











GROUP I 








Apparently stable 














40 


122 


19 







1 


9 


9 


41 


131 


26 







1 








42 


135 


31 







1 


6 





43 


136 


32 







1 


9 


9 


44 


500 


34 







1 


9 


9 


Lesser threat 
















45 


108 


2 


2 





1 





5 


46 


109 


3 


2 





1 





1 



APPENDIX 3 255 



Record Cavate 
Observation Number Number 



Amount Excavated 
Stability Tuff Type vs. Amount Natural 



Human 


Natural 


Damage 


Damage 





6 


2 


6 





6 





6 


9 


9 


9 


9 


9 


9 


9 


9 





6 





6 


9 


9 


9 


9 


9 


9 



47 


113 


7 


48 


114 


8 


49 


117 


12 


50 


124 


17 


Greater threat 






51 


111 


5 


52 


116 


10 


53 


120 


13 


54 


123 


14 


55 


132 


27 


56 


137 


33 


Major problem 






57 


125 


20 


58 


127 


22 


59 


140 


37 



2 
2 
2 
2 

3 
3 
3 
3 
3 
3 

4 
4 
4 









4 


3 








2 
1 
1 
2 

1 

2 
1 
2 
3 

2 

1 
1 

2 



irently stable 


1 




60 


8 


8 


61 


10 


10 


62 


23 


29 


63 


30 


34 


64 


36 


37 


65 


58 


57 


66 


56 


61 


67 


63 


65 


zr threat 






68 


1 


1 


69 


4 


4 


70 


13 


13 


71 


24 


30 



GROUP M 

4 
4 
4 
4 






4 
4 
4 



1 
1 
1 
2 
1 
1 
1 
1 

2 
1 
1 
1 



9 

9 
9 
9 


9 

9 

9 
9 



9 
1 
9 
9 
9 

3 
9 

9 
1 
9 
9 



256 CAVATE STRUCTURES 





Record 


Cavate 






Amount Excavated 


Human 


Natural 


Observation 


Number 


Number 


Stability 


Tuff Type 


vs. Amount Natural 


Damage 


Damage 


72 


25 


31 


2 





2 


9 


9 


73 


28 


35 


2 


4 


1 


9 


9 


74 


51 


51 


2 


4 


2 


9 


9 


75 


57 


58 


2 





2 





5 


76 


54 


59 


2 





2 





5 


77 


55 


60 


2 





1 


9 


9 


78 


61 


64 


2 














Greater threat 
















79 


2 


2 


3 


4 


2 


9 


9 


80 


14 


14 


3 


4 


2 


9 


9 


81 


64 


16 


3 


4 


2 





6 


82 


15 


18 


3 





1 


9 


9 


83 


17 


20 


3 


4 


2 


9 


9 


84 


29 


23 


3 


4 


2 





6 


85 


27 


24 


3 


1 


1 


9 


9 


86 


32 


33 


3 





1 





4 


87 


33 


36 


3 





1 


9 


9 


88 


34 


38 


3 


4 


2 


9 


9 


89 


40 


40 


3 





2 


9 


9 


90 


39 


44 


3 





2 


9 


9 


91 


42 


46 


3 





2 


9 


9 


92 


45 


48 


3 





2 


9 


9 


93 


46 


49 


3 





2 





6 


94 


47 


50 


3 


4 


2 





6 


95 


52 


52 


3 


4 


2 





6 


Major problem 
















96 


3 


3 


4 


4 


2 





4 


97 


9 


9 


4 


4 


2 


9 


9 


98 


65 


15 


4 


4 


2 





6 


99 


62 


17 


4 





2 


9 


9 


100 


41 


41 


4 





1 


9 


9 


101 


50 


55 


4 





1 


9 


9 



APPENDIX 3 



257 



Record Cavate Amount Excavated Human Natural 

Observation Number Number Stability Tuff Type vs. Amount Natural Damage Damage 



102 


53 


56 


4 



TSANKAWI 


2 





6 


Apparently stable 














103 


237 


28 




5 


1 





1 


104 


293 


35 




6 


1 





3 


105 


274 


52 







1 


9 


9 


106 


256 


58 




5 


1 





1 


107 


259 


59 







1 


9 


9 


108 


248 


64 




5 


1 


3 





109 


322 


67 







1 





1 


110 


241 


521 







1 


9 


9 


111 


265 


530 




5 


1 








112 


264 


531 







1 


9 


9 


113 


316 


564 




6 


1 








114 


503 


567 




6 


1 


9 


9 


115 


305 


568 




6 


1 


9 


9 


Lesser threat 
















116 


205 


15 


2 


5 


1 


9 


9 


117 


206 


16 


2 


5 


1 





5 


118 


207 


17 


2 


5 


2 


9 


9 


119 


208 


18 


2 


5 


1 





5 


120 


213 


19 


2 


5 


2 


5 


6 


121 


218 


21 


2 


5 


1 


9 


9 


122 


222 


24 


2 


5 


2 


9 


9 


123 


223 


25 


2 


5 


2 





5 


124 


216 


26 


2 





1 


5 


6 


125 


235 


27 


2 





2 


9 


9 


126 


289 


31 


2 


6 


2 





5 


127 


290 


32 


2 


6 


1 





5 


128 


291 


33 


2 


6 


1 


9 


9 



258 CAVATE STRUCTURES 





Record 


Cavate 




Observation 


Number 


Number 


Stabi 


129 


292 


34 


2 


130 


295 


37 


2 


131 


296 


38 


2 


132 


309 


40 


2 


133 


313 


41 


2 


134 


314 


42 


2 


135 


315 


43 


2 


136 


275 


51 


2 


137 


276 


53 


2 


138 


262 


55 


2 


139 


261 


56 


2 


140 


253 


61 


2 


141 


252 


63 


2 


142 


242 


65 


2 


143 


204 


501 


2 


144 


217 


508 


2 


145 


224 


509 


2 


146 


236 


517 


2 


147 


244 


523 


2 


148 


266 


533 


2 


149 


297 


553 


2 


150 


321 


570 


2 


Greater threat 








151 


285 


29 


3 


152 


288 


30 


3 


153 


294 


36 


3 


154 


319 


44 


3 


155 


320 


45 


3 


156 


267 


50 


3 


157 


263 


54 


3 


158 


254 


60 


3 


159 


247 


66 


3 



Amount Excavated Human Natural 
Stability Tuff Type vs. Amount Natural Damage Damage 



6 
6 
6 
6 
6 
6 
6 
5 

5 



5 
5 
3 
5 
5 


6 


6 
6 
6 
6 
6 







1 
2 



2 
3 
4 

2 

1 
2 

1 
1 

1 
2 
2 
1 
2 
1 
1 





9 
9 
9 
9 
9 

9 
1 

9 
9 





9 
9 
9 
9 

9 
9 
9 
9 
9 
9 
9 
9 
7 



6 
6 
9 
9 
9 
9 
9 

9 
5 
1 
9 
9 
3 
5 
1 
3 
6 
9 
9 
9 
9 

9 
9 
9 
9 
9 
9 
9 
9 




APPENDIX 3 259 



Record Cavate 
Observation Number Number 







Amount Excavated 


Human 


Natural 


bility 


Tuff Type 


vs. 


Amount Natural 


Damage 


Damage 


3 


5 




2 


9 


9 


3 


5 




3 





6 


3 







2 





6 


3 


6 




1 





6 


4 


5 




1 


5 


6 


4 


6 




1 


9 


9 


4 


4 




2 


9 


9 


4 










9 


9 



160 


219 


507 


161 


232 


514 


162 


233 


515 


163 


287 


548 


jor problem 






164 


221 


23 


165 


304 


39 


166 


239 


519 


167 


270 


536 



Note: Variable values and definitions can be found in appendix 1 . 



260 



CAVATE STRUCTURES 



Noncavate Stability Sorted by Rating, Group, and Number 

Record Number Cavate Number Stability Natural Damage Human Damage 



Apparently stable 




69 


3 


71 


6 


103 


36 


Lesser threat 




70 


5 


74 


9 


95 


27 


101 


33 


102 


35 


325 


42 


328 


45 


329 


46 


336 


53 


338 


55 


339 


56 


350 


59 


354 


74 


Greater threat 




83 


16 


91 


21 


92 


24 


93 


25 


94 


26 


96 


28 


98 


31 


77 


37 


105 


40 


326 


43 


327 


44 



GROUP A 

1 
1 
1 

2 
2 
2 
2 
2 
2 
2 
2 
2 
2 
2 
2 
2 

3 
3 
3 
3 
3 
3 
3 
3 
3 
3 
3 



1 
4 


2 
4 
6 
6 
5 
5 
6 
5 
6 
5 
2 
6 
4 

6 
6 
6 
6 
6 
6 
6 
6 
6 
6 
6 



1 



6 













1 














APPENDIX 3 261 



Record Number 


Cavate Number 


Stability 


Natural Damage 


Human Damage 


331 


48 


3 


6 





332 


49 


3 


6 





333 


51 


3 


6 





335 


52 


3 


6 





337 


54 


3 


6 


1 


341 


58 


3 


6 





344 


61 


3 


6 





349 


64 


3 


6 





348 


65 


3 


6 





355 


68 


3 


6 





356 


69 


3 


6 





353 


70 


3 


6 





351 


72 


3 


6 





Major problem 










72 


7 


4 


6 





73 


8 


4 


6 





79 


12 


4 


6 


7 


81 


14 


4 


6 


2 


88 


19 


4 


6 





78 


38 


4 


6 





324 


41 


4 


6 





84 


17 


9 
GROUP F 


9 


9 


Lesser threat 










143 


1 


2 


6 





148 


5 


2 


6 





153 


10 


2 


6 





158 


14 


2 


6 





163 


18 


2 


5 





165 


19 


2 


6 





164 


20 


2 


6 






262 CAVATE STRUCTURES 



Record Number 


Cavate Number 


Stability 


Natural Damage 


Human Damage 


169 


21 


2 


2 





172 


22 


2 


6 





179 


26 


2 


6 


1 


168 


29 


2 


6 





171 


30 


2 


6 





173 


32 


2 


6 





177 


33 


2 


6 





178 


34 


2 


6 





184 


39 


2 


6 





191 


41 


2 


6 





187 


44 


2 


2 





189 


46 


2 


6 





145 


47 


2 


6 





193 


50 


2 


6 





194 


52 


2 


6 





Greater threat 










146 


3 


3 


6 





149 


6 


3 


6 





150 


7 


3 


6 





151 


8 


3 


6 





154 


11 


3 


6 





156 


13 


3 


6 





162 


17 


3 


1 





176 


24 


3 


2 





175 


25 


3 


6 





170 


28 


3 


2 





182 


36 


3 


6 





185 


40 


3 


6 





190 


42 


3 


2 





188 


43 


3 


6 





157 


48 


3 


6 





203 


49 


3 


6 





192 


51 


3 


6 






APPENDIX 3 



263 



Record Number 


Cavate Number 


195 


53 


196 


54 


198 


55 


199 


57 


202 


61 


Major problem 




201 


58 


Apparently stable 




115 


9 


Lesser threat 




118 


16 


119 


18 


126 


21 


133 


28 


139 


35 


141 


36 


142 


38 


Greater threat 




107 


1 


110 


4 


112 


6 


121 


11 


128 


23 


129 


24 


130 


25 


134 


29 


Lesser threat 




26 


32 



Stability 



Natural Damage Human Damage 



2 
2 
2 
2 
2 
2 
2 

3 
3 
3 
3 
3 
3 
3 
3 



6 
6 
6 
6 
6 



6 
6 
6 
5 
6 
6 
2 

6 
6 
6 
6 
6 
6 
6 
6 











GROUP I 




















GROUP M 



264 CAVATE STRUCTURES 



Record Number 


Cavate Number 


Stability 


Natural Damage 


Human Damage 


Greater threat 










5 


5 


3 


2 





7 


7 


3 


9 


9 


19 


22 


3 


2 





21 


26 


3 


6 





31 


27 


3 


2 





35 


39 


3 


6 





43 


45 


3 


6 





44 


47 


3 


6 





66 


66 


3 


6 





Major problem 










6 


6 


4 


2 





11 


11 


4 


2 





12 


12 


4 


6 





22 


28 


4 


4 





48 


53 


4 


6 


1 


59 


62 


4 


6 





60 


63 


4 


6 



TSANKAWI 


Stable 










249 


525 




2 





250 


526 




6 





251 


527 




6 





257 


528 




6 





278 


541 




6 





302 


556 




5 





Lesser threat 










220 


22 


2 


6 





210 


503 


2 


6 





211 


504 


2 


6 





215 


505 


2 


6 


5 


214 


506 


2 


6 


5 



APPENDIX 3 265 



Record Number 


Cavate Number 


Stability 


Natural Damage 


Human Damage 


225 


510 


2 


5 





226 


511 


2 


1 





227 


512 


2 


5 





231 


513 


2 


6 





234 


516 


2 


6 





238 


518 


2 


6 





243 


522 


2 


2 





246 


524 


2 


5 





265 


532 


2 


6 


5 


269 


535 


2 


6 





272 


538 


2 


1 





273 


539 


2 


2 





277 


540 


2 


6 





279 


542 


2 


6 





280 


543 


2 


6 





281 


544 


2 


6 





283 


545 


2 


6 





282 


546 


2 


6 





284 


547 


2 


5 





288 


549 


2 


6 





298 


551 


2 


5 





299 


552 


2 


2 





303 


557 


2 


1 





306 


558 


2 


5 





307 


559 


2 


5 





308 


560 


2 


5 





310 


561 


2 


5 





311 


562 


2 


5 





312 


563 


2 


6 





317 


565 


2 


5 





318 


566 


2 


6 





323 


569 


2 


6 






266 



CAVATE STRUCTURES 



Record Number 


Cavate Number 


Stability 


Natural Damage 


Human Damage 


Greater threat 










255 


62 


3 


6 





209 


502 


3 


6 





260 


529 


3 


6 





268 


534 


3 


6 





271 


537 


3 


6 





286 


550 


3 


6 





300 


554 


3 


6 





Major problem 










258 


57 


4 


2 


4 



Note: Variable values and definitions can be found in appendix 1. 



Appendix 4: Detailed Listing of Rock Art 
June Crowder 

Summary of Rock Art Tables by Cavate Group 

Group A 

Number of cavates in Group A containing the indicated rock art: 



Abstracts 


8 


Cross 


1 


Geometries 


4 


Geometric abstracts 




Modern graffiti 


11 


Realistic bird 




Realistic snake 




Zig-zags 




Pictographs 


4 


Total 


32 


s of the individual cavates: 


Petroglyphs: 




Cavate 1 


modern graffiti 


2 


abstracts 


3 


geometric (enclosed hourglass) 


5 


modern graffiti 


10 


zig-zags 


12 


modern graffiti 


13 


abstracts, geometries, modern graffiti 


14 


modern graffiti 


15 


abstracts, realistic bird 


16 


abstracts, crosses (2), geometries 


18 


modern graffiti 


22 


abstracts, modern graffiti 


32 


abstracts, modern graffiti, realistic snake 


50 


abstracts, geometries, modern graffiti 


60 


abstracts, modern graffiti 


? 


(no number assigned; 2-3 cavates east of #75) abstracts, complex 




geometric abstract, modern graffiti 



Pictographs, or possible remains of pictographs: 

Cavate 15 small group of red splotches 

50 well executed red circle outlined in black, lower one-quarter and 

interior missing 
? (1-2 cavates east of #73) 4 small white stripes, lower one-quarter of 
wall has remnants of a white border and a small vertical stripe with 
horizontal slashes 



267 



268 CAVATE STRUCTURES 



? (2-3 cavates east of #73) small red stripes and red splotches 

A73 large white figure, white stars, etc. 



Group F 

Number of cavates on Group F containing the indicated rock art: 



Abstracts 


4 


Ceremonial figure 
Geometries 


2 
3 


Geometric abstracts 


1 


Human figure 
Masks 


1 
2 


Modern graffiti 
Realistic snake 


3 

1 


Stylized bird 
Stylized insect 
Terrace 


1 
1 
1 


Pictographs 


1 



Total 



21 



Contents of the individual cavates: 



Petroglyphs: 

Cavate 2 

24 
25 
26 
27 
31 
38 



Pictograph: 

Cavate 38 



abstracts, geometries, complex stylized bird, terraces, ceremonial 

figure 

simple mask, stylized mask, stylized insect (?) 

geometric abstract 

modern graffiti 

complex curvilinear abstract, realistic snakes (4), geometries, abstracts 

abstracts, modern graffiti 

ceremonial figure, modern graffiti, abstracts, geometric designs, 

human figure (armless), masks (4) 



possible remnants of pictographs composed of red paint splotches with 
superimposed incised lines 



APPENDIX 4 269 



Group I 

Number of cavates in Group I containing the indicated rock art: 

Abstracts 3 

Anthropomorphic figure 1 

Ceremonial figure 2 

Geometries 3 

Geometric abstracts 2 

Mask 

Modern graffiti 

Realistic animal 

Stick figure 

Stylized parrots 

Pictograph 

Total 17 

Contents of the individual cavates: 



Petroglyphs: 

Cavate 8 
12 
19 

26 



Pictographs: 

Cavate 19 



abstracts, geometries, ceremonial figures (3), stick figure 

geometric abstract 

stylized parrots (7), ceremonial figures (3), abstracts, geometries, 

modern graffiti 

geometries, abstracts, geometric abstract, squirrel profile, 

anthropomorph, masks (2) 



white circle (23 cm diameter), possible white parrot-type beak, two 
white solid rectangles, tan boxes (2) 



270 CAVATE STRUCTURES 



Group M 

Number of cavates in Group M containing the indicated rock art: 



Abstracts 


4 


Geometries 


3 


Human figure 
Hunter 


1 
1 


Masks 


3 


Realistic animal 


1 


Stylized birds 
Stick figure 
Zig-zags 
Pictograph 


1 
1 
1 
3 



Total 



Contents of the individual cavates: 



19 



Petroglyphs: 

Cavate 13 
18 
33 
40 
41 

60 

Pictographs: 

Cavate 13 
33 



41 



large zig-zags, abstracts 

abstracts geometries 

geometries, abstracts, stylized birds (4), mask 

profile walking-stick figure 

deer profile, hunter, small masks (9), geometric, outline of human 

torso (no features) 

abstracts, geometries, stylized mask, simple masks (3) 



white paint outline of right hand 

several design remnants composed of a partial red headdress, red 

stripes, yellow stripes, red and yellow paint, yellow paint with red 

dots, small black and yellow interlocking fret, some superimposition 

of incised lines 

red vertical realistic snake with superimposed incised lines 



APPENDIX 4 271 



Tsankawi (LA 50976) 

Number of cavates in LA 50976 containing the indicated rock art: 



Abstracts 


7 


Anthropomorphic figures 
Bird, realistic 


6 

2 


Ceremonial figures 
Cross 


2 

1 


Flute player 
Geometries 


2 
2 


Masks 


1 


Modern graffiti 
Realistic snake 


3 

2 


Serpent 
Serpent motif 
Serpent, two-horned 
Snake motif 


1 
2 
2 
2 


Terrace 


1 


Pictograph 


1 



Total 



Contents of the individual cavates: 



37 



Petroglyphs: 

Cavate 16 
20 
26 
33 
40 
41 
53 

54 
59 



61 
64 

66 



Pictographs: 

Cavate 59 



flute player, anthropomorph, abstracts 
anthropomorph, ceremonial figure, two-horned serpent 
anthropomorphic figure 
circle and dots on ceiling, abstracts 
outside entrance: 2 realistic birds, wavy line 
abstracts 

two-horned serpents (2), two-horned anthropomorph, head of two- 
horned serpent 

unfinished flute player, modern graffiti, abstracts 
cross, serpent motifs, abstracts, geometric, concentric circles, realistic 
bird figures (2), ceremonial figures (4), realistic snake, masks (4), 
terrace, snake motif 
ceiling snake motif 

abstracts, anthropomorph, serpents (one with a rattle), modern graffiti, 
geometries 

geometries, abstracts, serpent motifs, realistic snake, modern graffiti, 
anthropomorphic figure 



five small white stripes 



272 CAVATE STRUCTURES 

Cliff-face Petroglyphs 

Type and location of the various petroglyphs found on the cliff faces at Frijoles canyon and 
Tsankawi: 

Group A anthropomorphic figure, geometric design, bird, concentric circle, terrace 

Group I birds, snake, quadrupeds 

LA 50976 corn symbol, quadrupeds, anthropomorphic figures, birds, abstracts, arrow, snake 
motif, Kokopelli, sun symbol 



APPENDIX 4 273 



Historical Correlation with Chapman 

Twenty-three of Chapman's drawings (Hewett 1938) from Frijoles Canyon were matched with 
the originals during this survey. The table below lists the location of these specific petroglyphs. Also 
included is a brief description of noticeable changes from the drawings. 

Chapman (Hewett 19381 Location 

Plate III h 1-19 (photo 8) 

Plate IV f A- 13 (photo 2)— slight loss of bottom design due to 

deterioration of wall plaster 
k F-2 (photo 4) 

k F-38 (photo 4)~severe deterioration of middle design 

area due to loss of wall plaster 

Plate V a F-2 (photo 4) 

c A- 13 (photo 5) 

Plate VII a 1-19 (photo 9)~head shows severe deterioration due to 

loss of wall plaster 
b 1-19 (photo 3)-severe deterioration of body design due 

to loss of wall plaster, head marred with an incised X 
e 1-19 (photo 9) 

Plate VIII a 1-19 (photo 6) 

g 1-19 (photo 7) 

h 1-19 (photo 5) 

Plate IX a M-33 (photo 3) 

c M-33 (photo 7) 

d M-33 (photo 4) 

f M-33 (photo 3) 

h M-33 (photo 3)~slight deterioration of edge of body 

design due to loss of wall plaster 

Plate X a F-2 (photo 3)~slight deterioration of design due to loss 

of wall plaster 

Plate XI c 1-19 (photo 2)~incised X across face 

Plate XII c F-2 (photo 5) 

Plate Xin d 1-19 (photo 8) 

f M-60 (photo 2) 

j F-38 (photo 3A)-incised lines through smallest mask 



Appendix 5: Chamber Cluster Membership 
Chamber Cluster Analysis Listings 

Cluster Analysis Results Sorted by Cluster and Distance from Seed 



Cavate 


Cluster 


Volume 


Feature n 


Plaster 


From Seed 


TS 


53 


1 


4.83 


39 


2 


4.3 


TS 


55 


1 


6.14 


43 


1 


4.3 


TS 


50 


2 


9.77 


29 


3 


3.7 


TS 


26 


2 


9.47 


20 





4.0 


TS 


27 


2 


13.15 


43 


2 


4.1 


TS 


64 


2 


11.75 


50 


3 


4.4 


TS 


15 


2 


10.46 


27 





5.1 


TS 


66 


3 


17.00 


58 


4 


3.3 


TS 


20 


3 


15.50 


41 


3 


6.4 


TS 


59 


3 


18.63 


64 


2 


6.9 


M 


59 


4 


7.11 


36 


6 


1.9 


M 


35 


4 


2.72 


28 


7 


5.1 


A 


50 


4 


6.42 


32 


7 


5.2 


A 


18 


4 


6.17 


36 


3 


9.4 


TS 


16 


5 


4.79 


24 


3 


2.9 


F 


31 


5 


6.26 


19 


5 


3.1 


F 


35 


5 


3.51 


21 


4 


3.4 


A 


22 


5 


5.18 


26 


3 


3.7 


M 


38 


5 


5.65 


18 


2 


3.8 


TS 


30 


5 


5.93 


29 


2 


3.8 


A 


13 


5 


5.33 


27 


5 


4.4 


F 


45 


5 


4.05 


18 


6 


4.6 


F 


23 


5 


3.56 


25 


3 


4.9 


A 


32 


5 


4.07 


24 


7 


5.0 



275 



276 CAVATE STRUCTURES 

Cavate Cluster Volume Feature n Plaster From Seed 

5 7.58 21 7 5.6 

5 7.47 25 8 6.4 

5 4.78 29 9 7.6 

6 .21 0.8 
6 2.22 5 1 0.8 
6 1.59 3 1 0.9 
6 2.83 3 1.1 
6 1.29 5 1 1.2 
6 1.44 6 1 1.2 
6 1.82 4 1.3 
6 .20 1.4 
6 .40 1.4 
6 .40 1.4 
6 1.02 2 1 1.4 
6 0.93 5 1 1.5 
6 1.32 6 2 1.5 
6 3.43 5 1 1.5 
6 1.17 4 1.6 
6 3.12 6 1.6 
6 1.20 2 1.7 
6 2.32 4 1.7 
6 3.06 4 1.7 
6 0.45 2 1 1.8 
6 0.48 2 1 1.8 
6 0.56 3 1 1.8 
6 0.60 3 1 1.8 
6 0.86 2 1.8 
6 1.49 5 1 1.8 
6 0.59 8 1 1.9 
6 0.72 4 1.9 



A 


67 


A 


47 


F 


38 


I 


5 


I 


20 


TS 


501 


TS 


515 


I 


7 


M 


51 


TS 


19 


I 


3 


TS 


509 


TS 


517 


M 


37 


A 


30 


TS 


521 


M 


2 


TS 


519 


TS 


536 


F 


56 


A 


39 


I 


14 


F 


12 


M 


10 


A 


4 


TS 


508 


TS 


570 


M 


61 


TS 


531 


A 


63 



APPENDIX 5 277 



Cavate 


Cluster 


Volume 


Feature n 


Plaster 


From Seed 


TS 


533 


6 


0.73 


6 





1.9 


TS 


37 


6 


0.83 


5 





1.9 


I 


17 


6 


2.76 


6 


1 


1.9 


I 


32 


6 


2.85 


7 


1 


1.9 


F 


15 


6 


0.35 


5 


1 


2.0 


TS 


507 


6 


0.45 


6 


. 


2.0 


A 


57 


6 


0.64 


4 





2.0 


M 


8 


6 


3.26 


4 


1 


2.1 


M 


29 


6 


0.34 


5 


2 


2.2 


TS 


530 


6 


1.22 


5 





2.2 


TS 


40 


6 


0.85 


6 





2.3 


TS 


548 


6 


0.99 


6 





2.3 


M 


34 


6 


1.35 


19 


5 


2.3 


A 


66 


6 


0.62 


7 





2.4 


M 


41 


6 


0.59 


9 


1 


2.5 


TS 


23 


6 


2.84 


9 





2.9 


TS 


523 


6 


5.11 


6 





2.9 


A 


10 


6 


2.65 


17 


1 


3.0 


M 


57 


6 


0.51 


8 


1 


3.1 


I 


33 


6 


5.23 


3 


1 


3.2 


A 


2 


6 


. 


14 


4 


3.4 


TS 


41 


6 


4.09 


15 


1 


3.4 


I 


2 


6 


. 


1 


1 


3.5 


M 


1 


6 


4.76 


5 


3 


3.6 


A 


73 


6 


2.27 


19 


3 


4.5 


F 


4 


6 


5.11 


11 


4 


4.7 


TS 


18 


6 


1.33 


12 





4.8 


I 


12 


6 


3.07 


8 


6 


5.3 


TS 


44 


6 


8.18 


4 


3 


6.4 


TS 


25 


6 


9.25 


14 





7.2 


M 


31 


6 


7.38 


7 


6 


7.3 



278 CAVATE STRUCTURES 



Cavate Cluster Volume Feature n Plaster From Seed 



7 15.15 28 1 2.7 

7 22 1 3.0 

8 3.04 18 1 3.5 

8 4.94 43 5 3.5 

9 3.21 12 1.9 
9 3.14 10 2.6 
9 5.49 16 2 2.7 
9 0.74 10 1 3.0 
9 0.86 8 1 3.0 
9 1.23 7 3.1 
9 4.00 16 1 3.1 
9 2.69 24 3.2 
9 13 2 3.3 
9 4.28 16 4 3.3 
9 2.48 17 2 3.6 
9 4.10 14 4 3.7 
9 1.32 21 3.8 
9 1.48 15 4 3.8 
9 3,17 25 3 3.9 
9 4.49 16 3.9 
9 3.02 20 3 4.0 
9 2.24 16 1 4.6 
9 4.35 22 2 4.6 
9 2.04 10 1 5.2 
9 2.86 14 6 5.2 
9 3.25 22 5.5 
9 0.94 23 . 5.7 
9 6.69 25 2 6.1 



TS 


24 


TS 


21 


M 


13 


I 


19 


TS 


34 


TS 


58 


F 


16 


I 


31 


I 


34 


TS 


63 


TS 


33 


TS 


56 


I 


30 


A 


34 


A 


71 


A 


23 


TS 


52 


M 


36 


F 


27 


TS 


17 


F 


2 


F 


37 


I 


26 


F 


59 


I 


37 


TS 


61 


TS 


51 


TS 


54 



APPENDIX 5 279 



Cavate Cluster Volume Feature n Plaster From Seed 



I 


13 


10 


2.86 


12 


1 


2.7 


A 


60 


10 


5.25 


20 


2 


3.3 


I 


22 


10 


1.52 


14 


1 


3.5 


M 


55 


10 


2.69 


20 


3 


3.7 


TS 


28 


10 


2.95 


29 





3.7 


TS 


65 


10 


5.08 


21 


1 


3.7 


M 


30 


10 


1.84 


10 


1 


3.8 


M 


44 


10 


1.54 


17 


3 


4.0 


M 


9 


10 


6.39 


23 


1 


4.0 


M 


58 


10 


2.87 


25 


1 


4.3 


M 


60 


10 


5.25 


27 


2 


4.6 


M 


40 


10 


3.81 


15 


6 


4.8 


M 


20 


10 


6.31 


24 


3 


4.8 


TS 


45 


10 


7.02 


14 





4.9 


TS 


29 


10 


9.17 


19 


1 


5.2 



280 CAVATE STRUCTURES 

Ouster Analysis Results Sorted by Group and Cavate 

Cavate Cluster Volume Feature n" Plaster From Seed 

6 14 4 3.4 

6 0.56 3 1 1.8 

6 2.65 17 1 3.0 

5 5.33 27 5 4.4 

4 6.17 36 3 9.4 

5 5.18 26 3 3.7 
9 4.10 14 4 3.7 

6 0.93 5 1 1.5 

5 4.07 24 7 5.0 

9 4.28 16 4 3.3 

6 2.32 4 1.7 

5 7.47 25 8 6.4 

4 6.42 32 7 5.2 

6 0.64 4 2.0 

10 5.25 20 2 3.3 
6 0.72 4 1.9 
6 0.62 7 2.4 

5 7.58 21 7 5.6 
9 2.48 17 2 3.6 

6 2.27 19 3 4.5 

9 3.02 20 3 4.0 

6 5.11 11 4 4.7 

6 0.45 2 1 1.8 

6 0.35 5 1 2.0 

9 5.49 16 2 2.7 

5 3.56 25 3 4.9 

9 3.17 25 3 3.9 

5 6.26 19 5 3.1 

5 3.51 21 4 3.4 



A 


2 


A 


4 


A 


10 


A 


13 


A 


18 


A 


22 


A 


23 


A 


30 


A 


32 


A 


34 


A 


39 


A 


47 


A 


50 


A 


57 


A 


60 


A 


63 


A 


66 


A 


67 


A 


71 


A 


73 


F 


2 


F 


4 


F 


12 


F 


15 


F 


16 


F 


23 


F 


27 


F 


31 


F 


35 



APPENDIX 5 281 





Cavate 


Cluster 


Volume 


Feature n a 


Plaster 


From Seed 


F 


37 


9 


2.24 


16 


1 


4.6 


F 


38 


5 


4.78 


29 


9 


7.6 


F 


45 


5 


4.05 


18 


6 


4.6 


F 


56 


6 


1.20 


2 





1.7 


F 


59 


9 


2.04 


10 


1 


5.2 




2 


6 




1 


1 


3.5 




3 


6 


. 


2 





1.4 




5 


6 


. 


2 


1 


0.8 




7 


6 


1.29 


5 


1 


1.2 




12 


6 


3.07 


8 


6 


5.3 




13 


10 


2.86 


12 


1 


2.7 




14 


6 


3.06 


4 





1.7 




17 


6 


2.76 


6 


1 


1.9 




19 


8 


4.94 


43 


5 


3.5 




20 


6 


2.22 


5 


1 


0.8 




22 


10 


1.52 


14 


1 


3.5 




26 


9 


4.35 


22 


2 


4.6 




30 


9 


. 


13 


2 


3.3 




31 


9 


0.74 


10 


1 


3.0 




32 


6 


2.85 


7 


1 


1.9 




33 


6 


5.23 


3 


1 


3.2 




34 


9 


0.86 


8 


1 


3.0 




37 


9 


2.86 


14 


6 


5.2 


M 


1 


6 


4.76 


5 


3 


3.6 


M 


2 


6 


3.43 


5 


1 


1.5 


M 


8 


6 


3.26 


4 


1 


2.1 


M 


9 


10 


6.39 


23 


1 


4.0 


M 


10 


6 


0.48 


2 


1 


1.8 


M 


13 


8 


3.04 


18 


1 


3.5 



282 CAVATE STRUCTURES 





Cavate 


Cluster 


Volume 


Feature n a 


Plaster 


From Seed 


M 


20 


10 


6.31 


24 


3 


4.8 


M 


29 


6 


0.34 


5 


2 


2.2 


M 


30 


10 


1.84 


10 


1 


3.8 


M 


31 


6 


7.38 


7 


6 


7.3 


M 


34 


6 


1.35 


19 


5 


2.3 


M 


35 


4 


2.72 


28 


7 


5.1 


M 


36 


9 


1.48 


15 


4 


3.8 


M 


37 


6 


1.02 


2 


1 


1.4 


M 


38 


5 


5.65 


18 


2 


3.8 


M 


40 


10 


3.81 


15 


6 


4.8 


M 


41 


6 


0.59 


9 


1 


2.5 


M 


44 


10 


1.54 


17 


3 


4.0 


M 


51 


6 


1.44 


6 


1 


1.2 


M 


55 


10 


2.69 


20 


3 


3.7 


M 


57 


6 


0.51 


8 


1 


3.1 


M 


58 


10 


2.87 


25 


1 


4.3 


M 


59 


4 


7.11 


36 


6 


1.9 


M 


60 


10 


5.25 


27 


2 


4.6 


M 


61 


6 


1.49 


5 


1 


1.8 


TS 


15 


2 


10.46 


27 





5.1 


TS 


16 


5 


4.79 


24 


3 


2.9 


TS 


17 


9 


4.49 


16 





3.9 


TS 


18 


6 


1.33 


12 





4.8 


TS 


19 


6 


1.82 


4 





1.3 


TS 


20 


3 


15.50 


41 


3 


6.4 


TS 


21 


7 


. 


22 


1 


3.0 


TS 


23 


6 


2.84 


9 





2.9 


TS 


24 


7 


15.15 


28 


1 


2.7 


TS 


25 


6 


9.25 


14 





7.2 


TS 


26 


2 


9.47 


20 





4.0 



APPENDIX 5 283 

Cavate Cluster Volume Feature n' Plaster From Seed 

2 13.15 43 2 4.1 

10 2.95 29 3.7 

10 9.17 19 1 5.2 

5 5.93 29 2 3.8 
9 4.00 16 1 3.1 

9 3.21 12 1.9 

6 0.83 5 1.9 
6 0.85 6 2.3 
6 4.09 15 1 3.4 
6 8.18 4 3 6.4 

10 7.02 14 4.9 

2 9.77 29 3 3.7 
9 0.94 23 5.7 
9 1.32 21 3.8 
1 4.83 39 2 4.3 
9 6.69 25 2 6.1 

1 6.14 43 1 4.3 
9 2.69 24 3.2 
9 3.14 10 2.6 

3 18.63 64 2 6.9 
9 3.25 22 5.5 

9 1.23 7 3.1 

2 11.75 50 3 4.4 

10 5.08 21 1 3.7 

3 17.00 58 4 3.3 
6 1.59 3 1 0.9 
6 0.45 6 2.0 
6 0.60 3 1 1.8 
6 4 1.4 
6 2.83 3 1.1 
6 4 1.4 



TS 


27 


TS 


28 


TS 


29 


TS 


30 


TS 


33 


TS 


34 


TS 


37 


TS 


40 


TS 


41 


TS 


44 


TS 


45 


TS 


50 


TS 


51 


TS 


52 


TS 


53 


TS 


54 


TS 


55 


TS 


56 


TS 


58 


TS 


59 


TS 


61 


TS 


63 


TS 


64 


TS 


65 


TS 


66 


TS 


501 


TS 


507 


TS 


508 


TS 


509 


TS 


515 


TS 


517 



284 CAVATE STRUCTURES 





Cavate 


Cluster 


Volume 


Feature n a 


Plaster 


From Seed 


TS 


519 


6 


1.17 


4 





1.6 


TS 


521 


6 


1.32 


6 


2 


1.5 


TS 


523 


6 


5.11 


6 





2.9 


TS 


530 


6 


1.22 


5 





2.2 


TS 


531 


6 


0.59 


8 


1 


1.9 


TS 


533 


6 


0.73 


6 





1.9 


TS 


536 


6 


3.12 


6 





1.6 


TS 


548 


6 


0.99 


6 





2.3 


TS 


570 


6 


0.86 


2 





1.8 



"Feature n is all features except walls. Not all of the features in the count were included in the cluster 
analysis. 



References 



Bailey, R. A., R. L. Smith, and C. S. Ross 

1969 Stratigraphic Nomenclature of Volcanic 
Rocks in the Jemez Mountains, New Mexico. 
U.S. Geological Survey Bulletin 1274-P. 
U.S. Government Printing Office, 
Washington. 

Bandelier, Adolph F. 

1971 The Delight Makers. Harcourt Brace 
Jovanovich, New York. Originally published 
1890. 

Beam, George L. 

1906 The Prehistoric Ruin of Tsankawi. 
National Geographic 20:807-822. 

Bierbower, Susan 

1905 Among the Cliff and Cavate Dwellings 
of New Mexico. Records of the Past 4:226- 
233. 

Blair, Jonathan S. 

1970 Keeping House in a Cappadocian Cave. 
National Geographic 138:126-146. 

Breternitz, David A. 

1966 An Appraisal of Tree-Ring Dated Pottery 
in the Southwest. Anthropological Papers of 
the University of Arizona 10. University of 
Arizona, Tucson. 

Canby, Thomas Y. 

1982 The Anasazi: Riddles in the Ruins. 
National Geographic 162:562-592. 

Carlson, Ingrid K., and T. A. Kohler 

1989 LA 50972 (Cavate M77). In Bandelier 
Archaeological Excavation Project: Research 



Design and Summer 1988 Sampling, edited 
by T. A. Kohler, pp. 50-59. Reports of 
Investigations 61. Washington State 
University, Pullman. 

Caywood, Louis R. 

1966 The Excavation of Rainbow House, 
Bandelier National Monument, New Mexico. 
National Park Service, Southwestern 
Archeological Center, Globe, Ariz. 



Chapman, Kenneth M. 

1916 Graphic Art of the Cave Dwellers. 
Palacio 3 (2):37-41. 



El 



1938 The Cave Pictographs of the Rito de los 
Frijoles. In Pajarito Plateau and Its Ancient 
People, by E. L. Hewett. University of New 
Mexico Press, Albuquerque. 

Colton, Harold S. 

1932 A Survey of Prehistoric Sites in the 
Region of Flagstaff, Arizona. Smithsonian 
Institution Bureau of American Ethnology 
Bulletin 104. Washington, D.C. 

1946 The Sinagua: A Summary of the 
Archaeology of the Region of Flagstaff, 
Arizona. Museum of Northern Arizona 
Bulletin No. 22. Flagstaff, Arizona. 

Cordell, Linda S. 

1979 A Cultural Resources Overview of the 
Middle Rio Grande Valley, New Mexico. 
USDI Bureau of Land Management and 
USDA Forest Service, Santa Fe and 
Albuquerque. 



285 



286 CAVATE STRUCTURES 



Dane, Carle H., and George O. Bachman 

1965 Geologic Map of New Mexico. U.S. 
Geologic Survey, Washington, D.C. 

Dougherty, Julia D. 

1980 An Archeological Evaluation of Tsiping 
Ruin. Cultural Resources Report 1, Santa Fe 
National Forest. USDA, Forest Service, 
Santa Fe. 

Douglass, William Boone 

1917 Notes on the Shrines of the Tewa and 
Other Pueblo Indians of New Mexico. 
Proceedings of the XIX Congress of 
Americanists, pp. 344-378. Washington, 
D.C. 

Fewkes, J. Walter 

1904 Two Summers' Work in Pueblo Ruins. 
In 22nd Annual Report of the Bureau of 
American Ethnology, pp. 3-195. Government 
Printing Office, Washington, D.C. 

1910 The Cave Dwellings of the Old and New 
Worlds. American Anthropologist 12:390- 
416. 

1913 Antiquities of the Upper Verde and 
Walnut Creek Valleys, Arizona. In 28th 
Annual Report of the Bureau of American 
Ethnology, pp. 181-220. Government 
Printing Office, Washington, D.C. 

Fliedner, Dietrich 

1975 Pre-Spanish Pueblos in New Mexico. 
Annals of the Association of American 
Geographers 65:363-377. 

Hall, Susan D. 

1992 An Architectural Analysis of Cavate 
Dwellings in the Verde Valley, Arizona. 
M.A. thesis, Northern Arizona University, 
Flagstaff. 

Harrington, John P. 

1916 Ethno geography of the Tewa Indians. 



Bureau of American Ethnology, 29th Annual 
Report. Smithsonian Institution, 

Washington, D.C. 



Heiken, Grant 

1979 Pyroclastic Flow Deposits. 
Scientist 67 (5):564-571. 



American 



Hendron, J. W. 

1940 Prehistory of El Rito de los Frijoles: 
Bandelier National Monument. Technical 
Series 1. Southwestern Monuments 

Association, Coolidge, Ariz. 

1943 Group M of the Cliff Dwellings, Rooms 
1, 2, 3, 4, and 5, Caves 1, 2, 3, and 4, 
Frijoles Canyon, Bandelier National 
Monument, New Mexico. Ms. on file, 
Bandelier National Monument. 

1946 We Found America's Oldest Tobacco. 
New Mexico Magazine 24 (11): 11-13, 33-37. 

Hewett, Edgar L. 

1908 Excavations at Puye in 1907. Papers of 
the School of American Archaeology 4. Also 
published as Archaeology of Rio Grande 
Valley, Out West 31, no. 2 (August 1909). 

1909a The Excavations at Tyuonyi, New 
Mexico, in 1908. American Anthropologist 
11:434-455. 

1909b Excavations at El Rito de los Frijoles in 
1909. American Anthropologist 1 1:651 -673 . 

1938 Pajarito Plateau and Its Ancient People. 
University of New Mexico Press, 
Albuquerque. 

Hill, James N., and W. Nicholas Trier weiler 
1986 Prehistoric Response to Food Stress on 
the Pajarito Plateau, New Mexico: Technical 
Report and Results of the Pajarito 
Archaeological Research Project, 1977-1985. 
National Science Foundation Grant 



REFERENCES 287 



BNS-78-08118. Department of Anthro- 
pology, University of California at Los 
Angeles. 

Hub bell, Lyndi, and Diane Tray lor 

1982 Bandelier: Excavations in the Flood Pool 
ofCochiti Lake, New Mexico. National Park 
Service, Southwest Cultural Resources 
Center, Santa Fe. 

Hyland, Justin R. 

1986 A Preliminary Investigation of a 
Prehistoric Cavate Site on the Pajarito 
Plateau, North Central New Mexico. Honors 
thesis, Department of Anthropology, 
University of California, Berkeley. 

Johnson, Chester 

1960 Report on the Excavation of a Child's 
Burial in Cave Room C54, Tsankawi Ruin, 
Bandelier National Monument. Ms. on file at 
Bandelier National Monument. 

Kelley, V. C, E. H. Baltz, and R. A. Bailey 

1961 Road Log: Jemez Mountains and 
Vicinity. In New Mexico Geological Society 
Twelfth Field Conference, pp. 47-60. New 
Mexico Bureau of Mines and Mineral 
Resources, Socorro. 

Kempe, David 

1988 Living Underground: A History of Cave 
and Cliff Dwelling. The Herbert Press, 
London. 

Kent, Kate P. 

1983a Prehistoric Textiles of the Southwest. 
University of New Mexico Press, 
Albuquerque. 

1983b Pueblo Indian Textiles: A Living 
Tradition. School of American Research 
Press, Santa Fe. 



Kern, Willis F., and James R. Bland 

1934 Solid Mensuration. John Wiley 
Sons, New York. 



and 



Kohler, Timothy A., and Angela R. Linse 

1993 Papers on the Early Classic Period 
Prehistory of the Pajarito Plateau, New 
Mexico. Reports of Investigations 65. 
Department of Anthropology, Washington 
State University, Pullman. 



Lang, David M. 

1966 The Georgians. 



Praeger, New York. 



Lang, Richard W. 

1982 Transformation in White Ware Pottery 
of the Northern Rio Grande. In Southwestern 
Ceramics: A Comparative Overview, edited 
by A. H. Schroeder, pp. 153-200. Arizona 
Archaeologist 15. Arizona Archaeological 
Society, Phoenix. 

Lange, Charles H., and Carroll L. Riley 

1966 The Southwestern Journals of Adolph F. 
Bandelier, 1880-1882. University of New 
Mexico Press, Albuquerque. 

Lange, Charles H., Carroll L. Riley, and 
Elizabeth M. Lange 

1975 The Southwestern Journals of Adolph F. 
Bandelier, 1885-1888. University of New 
Mexico Press, Albuquerque. 

Lekson, Stephen H. 

1988 The Idea of the Kiva in Anasazi 
Archaeology. The Kiva 53:213-234. 

Lent, Stephen C. 

1988 Excavations of a Late Archaic Pit 
Structure (LA 51912), near Otowi, San 
Ildefonso Pueblo, New Mexico. Lab- 
oratory of Anthropology Notes 52. Museum 



288 CAVATE STRUCTURES 



of New Mexico, Office of Archaeological 
Studies, Santa Fe. 

Lister, Robert H. 

1939 Stabilization of Long House, Bandelier 
National Monument. Ms. on file, Bandelier 
National Monument. 

1940a Stabilization of Frijoles Cave Ruins, 
1940, Bandelier National Monument. Ms. on 
file, Bandelier National Monument. 

1940b Stabilization of Cave Ruins, Bandelier 
National Monument, 1940. Ms. on file, 
Bandelier National Monument. 

McKenna, Peter J., and Robert P. Powers 

1986 Preliminary Report on the 1985 Test 
Survey of Bandelier National Monument, 
New Mexico. Ms. on file, Intermountain 
Cultural Resource Center, Anthropology 
Program, National Park Service, Santa Fe. 

Magers, Pamela C. 

1986 Weaving at Antelope House. In 
Archeological Investigations at Antelope 
House, compiled by D. P. Morris, pp. 224- 
276. National Park Service Publications in 
Archeology 19. National Park Service, 
Washington, D.C. 

Mathien, F. Joan, Charlie R. Steen, and 
Craig D. Allen 

1993 The Pajarito Plateau: A Bibliography. 
Southwest Regional Office, Division of 
Anthropology, Southwest Cultural Resources 
Center Professional Paper 49. Southwest 
Cultural Resources Center, Santa Fe. 

Maxon, James C. 

1962 A Cave Kiva in Bandelier National 
Monument. Ms. on file, Intermountain 
Cultural Resource Center, Anthropology 
Program, National Park Service, Santa Fe. 

1969 A Study of Two Prehistoric Pueblo Sites 
on the Pajarito Plateau of New Mexico. 



M. A. thesis, University of Wisconsin. 

Mera, Harry P. 

1932 Style Trends of Pueblo Pottery in the 
Rio Grande and Little Colorado Culture 
Areas in the Sixteenth to Nineteenth 
Centuries. Laboratory of Anthropology 
Technical Series Bulletin 4. Laboratory of 
Anthropology, Santa Fe. 

1934 A Survey of the Biscuit Ware Area in 
Northern New Mexico. Laboratory of 
Anthropology Technical Series Bulletin 6. 
Laboratory of Anthropology, Santa Fe. 

Mills, Barbara M. 

1986 Temporal Variability in the Ceramic 
Assemblages of the Eastern Slope of the 
Black Range, New Mexico. In Mogollon 
Variability, edited by C. Benson and S. 
Upham, pp. 169-180. University Museum 
Occasional Papers 15. New Mexico State 
University, Las Cruces. 

Mindeleff, Cosmos 

1896 Aboriginal Remains in Verde Valley, 
Arizona. In Thirteenth Annual Report of the 
Bureau of American Ethnology, by J. 
W. Powell, pp. 179-261. U.S. Government 
Printing Office, Washington, D.C. 

Morley, Sylvanus G. 

1910 The South House, Puye\ Papers of the 
School of American Archaeology, O.S. 7. 
Archaeological Institute of America. 

Onstott, Thomas B. 

1948 Excavations at Tyuonyi, Bandelier 
National Monument, 1947-48. Southwestern 
Monuments Library, Western Archaeological 
and Conservation Center, Tucson. 

Orcutt, Janet D. 

1994 Chronology. In The Bandelier 

Archaeological Survey. Edited by Robert P. 
Powers and Janet D. Orcutt. Ms. on file, 



REFERENCES 289 



Intermountain Cultural Resource Center, 
Anthropology Program, National Park 
Service, Santa Fe. 

Peckham, Stewart 

1979 When Is a Rio Grande Kiva? In 
Collected Papers in Honor of Bertha Pauline 
Dutton, edited by A. H. Schroeder, pp. 55- 
86. Papers of the Archaeological Society of 
New Mexico 4. Archaeological Society of 
New Mexico, Albuquerque. 

Powell, John Wesley 

1886 Annual Report of the Director. In 
Fourth Annual Report of the Bureau of 
Ethnology, 1882-1883. U.S. Government 
Printing Office, Washington, D.C. 

1891 Annual Report of the Director. In 
Seventh Annual Report of the Bureau of 
Ethnology, 1885-1886. U.S. Government 
Printing Office, Washington, D.C. 



Prudden, T. Mitchell 

1903 Ruins of the San Juan Watershed in 
Utah, Colorado, and New Mexico. American 
Anthropologist n.s. 5:224-288. 

Riboud, Marc 

1958 Cappadocia: Turkey's Country of Cones. 
National Geographic 113:122-146. 

Robinson, William J., John W. Hannah, and 
Bruce G. Harrill 

1972 Tree-Ring Dates from New Mexico I, O, 
U: Central Rio Grande Area. Laboratory of 
Tree-Ring Research, University of Arizona, 
Tucson. 

Ross, C. S., R. L. Smith, and R. A. Bailey 

1961 Outline of the Geology of the Jemez 
Mountains, New Mexico. In New Mexico 
Geological Society Twelfth Field Conference, 
pp. 139-143. New Mexico Bureau of Mines 
and Mineral Resources, Socorro. 



Powers, Robert P. 

1988 Final Archeological Research Design for 
a Sample Inventory of Bandelier National 
Monument. Ms. on file, Intermountain 
Cultural Resource Center, Anthropology 
Program, National Park Service, Santa Fe. 

Preucel, Robert W. 

1985 Preliminary Report of the Pajarito Field 
House Project: 1985 Season. Report 
prepared for the Santa Fe National Forest, 
Southwestern Division, Albuquerque. 

1986a The Pajarito Field House Project. 
Archaeology at UCLA 2 (19). 

1986b Preliminary Report of the Pajarito Field 
House Project: 1986 Season. Report 
prepared for the Santa Fe National Forest, 
Southwestern Division, Albuquerque. 

1987 Settlement Succession on the Pajarito 
Plateau, New Mexico. The Kiva 53 (l):3-33. 



SAS Institute 

1985 SAS User's Guide: Basics. Version 5 
edition. SAS Institute, Cary, N.C. 

Schaafsma, Polly 

1975 Rock Art in the Cochiti Reservoir 
District. Museum of New Mexico Papers in 
Anthropology 16. Museum of New Mexico, 
Santa Fe. 

1980 Indian Rock Art of the Southwest. 
University of New Mexico Press, 
Albuquerque. 



Severy, Merle 

1983 Gifts of Golden Byzantium. 
Geographic 164:722-737. 



National 



Siegel, Sidney 

1956 Nonparametric 
Behavioral Sciences. 
York. 



Statistics for the 
McGraw-Hill, New 



290 CAVATE STRUCTURES 



Smiley, Terah L. 

1951 A Summary of Tree-Ring Dates from 
Some Southwestern Archaeological Sites. 
Laboratory of Tree-Ring Research Bulletin 5. 
University of Arizona, Tucson. 

Smiley, Terah L., Stanley Stubbs, and Bryant 
Bannister 

1953 A Foundation for the Dating of Some 
Late Archaeological Sites in the Rio Grande 
Area, New Mexico: Based on Tree-Ring 
Methods and Pottery Analyses. Laboratory 
of Tree-Ring Research Bulletin 6. University 
of Arizona, Tucson. 

Smith, Watson 

1952 Excavations in Big Hawk Valley, 
Wupatki National Monument, Arizona. 
Museum of Northern Arizona Bulletin 24. 
Museum of Northern Arizona, Flagstaff. 

1972 Prehistoric Kivas of Antelope Mesa, 
Northeastern Arizona. Papers of the Peabody 
Museum of Archaeology and Ethnology 39, 
1. Harvard University, Cambridge. 

Snow, David H. 

1974 The Excavation of Saltbush Pueblo, 
Bandelier National Monument, New Mexico. 
Laboratory of Anthropology Note 97. 
Museum of New Mexico, Santa Fe. 

1982 The Rio Grande Glaze, Matte-Paint, and 
Plainware Tradition. In Southwestern 
Ceramics: A Comparative Overview, edited 
by A. H. Schroeder, pp. 235-278. Arizona 
Archaeologist 15. Arizona Archaeological 
Society, Phoenix. 

Steen, Charlie R. 

1977 Pajarito Plateau Archaeological Survey 
and Excavations. Los Alamos Scientific 
Laboratory LASL-77-4. 

1979 The Mortendad Style of Rock Art, 
Pajarito Plateau, New Mexico. In Collected 



Papers in Honor of Bertha Pauline Dutton, 
edited by A. H. Schroeder, pp. 41-53. 
Papers of the Archaeological Society of New 
Mexico 4. Archaeological Society of New 
Mexico, Albuquerque. 

1982 Pajarito Plateau Archaeological Surveys 
and Excavations, II. Los Alamos Scientific 
Laboratory LA-8860-NERP. Available from 
National Technical Information Service, U.S. 
Department of Commerce, Springfield, Va. 

Stuart, David E., and Rory P. Gauthier 

1981 Prehistoric New Mexico: Background for 
Survey. New Mexico Historic Preservation 
Bureau, Santa Fe. 

Thompson, Raymond H., editor 

1990 When Is a Kiva? and Other Questions 
about Southwestern Archaeology by Watson 
Smith. University of Arizona Press, Tucson. 

Turney, John F. 

1948 An Analysis of the Material Taken from 
a Section of Group M of the Cliffs, Frijoles 
Canyon, Bandelier National Monument, New 
Mexico, 1943. M.A. thesis, Adams State 
College, Alamosa, Colo. 

Van Zandt, Tineke R. 

1994 Architecture and Site Structure. In The 
Bandelier Archeological Survey. Edited by 
Robert P. Powers and Janet D. Orcutt. Ms. 
on file, Intermountain Cultural Resource 
Center, Anthropology Program, National 
Park Service. 

Warren, A. Helene 

1979 Ceramic Studies in Cochiti Reservoir, 
1976-1977. In Archaeological Investigations 
in Cochiti Reservoir, New Mexico, vol. 3, 
edited by J. Biella, pp. 27-42. Office of 
Contract Archaeology, University of New 
Mexico, Albuquerque. 



REFERENCES 291 



White, G. E. Whittaker, John C. 

1904 The Cavate Dwellings of Cappadocia. n.d. J. W. Fewkes's Photographs of New 

Records of the Past 3:66-73. Caves in 1900. Ms. submitted to Kiva. 



Index 



adobe collar, 149 

adobe ramp, 142. See also mealing 

complexes 
Alamo Canyon, 5 
American Anthropologist, 8 
archaeomagnetic dating, 62 
artifacts and collections, 9, 11, 85, 97, 236 
Awanyu, 8, 51; illustrated, 192, 194 
axe groove (axe-sharpening groove), 140, 

149, 235 

back wall features, 84, 161, and exterior 

openings, 122; in noncavates, 151, 157 
Bailey, R. A., R. L. Smith, and C. S. Ross, 

2, 15, 16, 53 
Bandelier, Adolph F., 5-6, 216 
Bandelier Black-on-gray, 11, 17 
Bandelier National Monument, 1, 2, 11, 17. 

See also Frijoles Canyon; Tsankawi 
Bandelier Tuff, 3, 9, 15, 16. See also tuff 

types 
Barthuli, K., and S. Hall, 49 
Barthuli, K., J. Vint, and W. Bustard, 47 
Beam, George L., 1, 7-8, 13 
beam features, as functional, 201; holes for 

beams/vigas (beam seat), 107, 157,165, 

182, 233; niches and beams, 203. See 

also ceilings 
beam seat, 165, 182, 233; viga holes, 107, 

157. See also beam features 
Bierbower, Susan, 1, 7 
Big Kiva, 29, 63 
bins, plank slots for, 157 
birds, illustrated designs for, 193 



Biscuit B ceramic, 64 

Blair, Jonathan S., 4, 94, 138 

boundary, cultural, 217; cultural context, 

16,59 
Bretemitz, David A., 67 
Bureau of American Ethnology, 6 
burials, 8, 11, 53 
burns, floor, 138; firepits, 138, 201, and 

deflectors, 150; hearths, 51, 62 

C-14 Sample, 64 

Camp Hamilton, 1 1 

Canada de Cochiti, 15 

Canby, Thomas Y., 13 

Cappadocia, Turkey, 4, 94, 134, 138 

Carlson, Ingrid K., and T. A. Kohler, 9, 
13, 51, 126, 157 

cavate type, 85; definitions, 1-2, 17, 78, and 
functional categories, 13; filled cavate, 
84, (number of) 108; partial cavate, 84, 
(number of) 108; variation in form, 77. 
See also chambers 

Cay wood, Louis R., 45 

ceilings, 134, 233; vertical holes in, 61, 
185, 235; viga holes in, 157, 161. See 
also beam features; plastering; roofing 

ceramic patterns, 72-75. See also ceramic 
sample 

ceramic sample, 10, 68-70, 73, 216; surface 
sherd counts, 9, 64; Turney's types, 9. 
See also ceramic patterns; ceramics 

ceramics, dating, 61, 63, for Group M, 45; 
form distributions, 72, 74, 75, and site 
function, 75; rock art dating with 



293 



294 CAVATE STRUCTURES 



ceramics, 196. See also ceramic patterns 

Ceremonial Cave, 27, 61 

Cerro Pedernal, 2 

Chama River, 15, as Rio Chama, 216 

chamber corner, 132, 233, 234 

chamber location, 80, 90, 231, 233, 234; 
levels (stories), 83, 98; rooms up/rooms 
down, 82. See also chambers 

chambers, defining, 84, 108, and back 
chambers, 150; combined chambers, 138, 
and expanded chambers, 59, 83; 
comparisons, 59, 117-20; functions, 13, 
112, 116, 213, and functional clustering, 
208-11; recording location, 82-83, and 
shape and size, 80-81, 91-93, 112, 208, 
226, 229, 232, 239; volume as 
functional, 201. See also cavate type; 
chamber location; and rooms separately 

Chapman, Kenneth M., 8, 17, 185, 190, 
197; survey of rock art by, 8, 273 

chronology, 5, 61, and dating with 
cei amies, 45, 61, 63, 169; Classic 
Period, 17, 62, 67, 71; Coalition Period, 
17, 45, 62, 67, 71, 71-72; 
Developmental Period, 71; Early Classic, 
216; Late Coalition, 216; Pueblo III, 5. 
See also occupation 

Civilian Conservation Corps (CCC), 45 

Classic Period (Rio Grande Classic), 17, 62, 
67,71 

cliff niche, 84, 190, 235 

cluster analysis, 206-11, 275 

Coalition Period, 17, 45, 62, 67, 71, 71-72 

Cochiti Pueblo, 2 

collections, 85, and artifacts, 9, 11, 97, 236 

Colton, Harold S., 5 

combined chamber, 138; expanded 
chambers, 59, 83 

compass location point, 138, 235 

computers, recording with (direct data 
entry), 87, and data manipulation, 90, 93 

condition/damage, 84, 99-105 

construction and use, 94-96; recording 
evidence for, 81, 226, 233; reoccupation, 
11, 74, and historic use, 7, 9, 71, 74. 
See also occupation 



Cordell, Linda S., 17 

Corral Hill, 71 

cotton, 216; weaving, 215. See also loom 

features 
Crowder, Bill, 86, 93 
Crowder, June, 190, 191; rock art study by, 

197, 267-72 
Crowder, June, and Bill Crowder, 85, 93 
Cuevitas Arribas, 20, 61 
culinary ware, 72, 74, 75 
cultural context, 16, and comparisons, 59; 

cultural boundaries, 217 

dado, incised, 189, 233; illustrated, 153, 

192 
damage, 84, 99-105, 231 
data manipulation, 90 
deflector, 150, 201, 234; illustrated, 152 
Delight Makers, The, Bandelier, 6 
Developmental Period, 71 
Die Koshare, Bandelier, 6 
digging sticks, 94 
direct data entry, 87 
doors, as functional features, 201, 229; 

exterior, 116, and exterior openings not 

doors, 122; interior, 122; niches as false 

doors, 190; passages, 138 
Dougherty, Julia D., 2 
Douglass, William Boone, 9 

Early Classic, 216 

economics, cotton in the, 216; the growing 

zone, 16 
Ellis, F. H., 65 
environment, 15 
ethnicity, canyon and mesa, 59; cultural 

boundary, 217; cultural context, 16 
exterior door, 116. See also doors 
exterior opening, 122, 233. See also doors 

fauna, 9, 45, 99, 227 

feature co-occurance, 201 

feature fraction, 81, 231 

feature types, 84, 228, 231, 233, 235, and 

canyon/mesa comparisons, 61. See also 

feature types separately 



INDEX 295 



Fewkes, J. Walter, 1-2, 4, 5, 9 

field houses, cavates as, 13, 215 

field time, 88-90 

FIELDLA, 90 

fill, 96, 98, 226; incised dado and, 190; 

filled cavate, 84, 108 
filled cavate, 84, and number of, 108 
firepits, 138, 201, and deflectors, 150; floor 

burns, 138, 201; hearths, 51, 62 
Flagstaff, Arizona, 4, 5 
Fliedner, Dietrich, 2, 215 
floor burn, 138, 201; firepits, 138, 201, and 

deflectors, 150; hearths, 51, 62 
floor depression, 140, 234 
floor features, 138-50, as functional, 201; 

floors, 134. See also floor features 

separately 
floor pit, 201 

floor pit complex, 140, 235 
floor ridge, 142, 201, and slots, 157 
floors, 134; floor features, 138-50, 201 
Four Corners Area, cavates of the, 4 
Frijoles Canyon, 14, 17-18, and 
comparisons, 59, 112, 116, 157, 197, 

217; Bandelier at, 5; Hewett at, 8; Lister 

at, 11; Stevenson at, 6. See also Group 

A; Group F; Group I; Group M; and see 

sites separately 
Frijolito site, 63 



140, 142, 152, 161, 185, 188, 189, and 
comparisons, 59, 61; other collections 
for, 85 

Group B, 71, 74; unrecorded rooms at, 132 

Group C, 71, 74 

Group D (Long House), 8, 11, 20, 216 

Group E (Snake Village and Sun House), 8 

Group F, ceramics and dating for, 62, 65, 
71, 74; identification for, 20, and 
setting, 29-39, 82, 216; rock art at, 192, 
197, 268; room plans, 33-35, 86, 91, 
94, and features, 94, 122, 132, 134, 
138, 149, 150, 161, 185, 189; 
unrecorded rooms for, 132 
Group I, ceramics and dating for, 62, 65, 
71, 74, 75; identification for, 20, and 
setting, 21, 37, 45^6, 82, 84, 216; rock 
art at, 197, 269, 272; room plans, 41, 
43, 86, 91, 94, and features, 45, 122, 
132, 134,149, 150 

Group J, mentioned, 37 

Group M, ceramics and dating for, 62, 63, 
65, 67, 71, 74; excavation at, 9, 13, 51, 
and collections, 85; identification for, 
20, and setting, 21, 45, 51, 52, 216; 
rock art at, 51, 192, 197, 270; room 
plans, 47, 49, 78, 84, 86, 94, and 
features, 51, 122, 126, 132, 138, 142, 
149, 157, 189, and comparisons, 61 



Garcia Canyon, 13, 62, 105, 116 

Gauthier, R., 62 

geology, 15, and topography compared, 59 

geometric fraction, 81, 230 

Glaze C ceramic, 71 

Glaze D polychrome ceramic, 71 

Glaze E ceramic, 71 

Greene, Ed, 52 

groove, 233; axe-sharpening, 140; wall 
groove, 185, 201 

group attributes, 93-94 

Group A, ceramics and dating for, 62, 64- 
65, 71; identification for, 20, and setting, 
20-28, 82, 84, 216, and deterioration, 
28-29; rock art at, 192, 197, 267, 272; 
room plans, 22-25, 86, 91, 94, and 
features, 29, 122, 126, 132, 134, 138, 



habitation rooms, 13, 206, 213, 227, and 

comparisons, 112 
Hall, Susan, 2, 4 
hand-and/or-toe holds, 84, 108, 189, 233; 

routes to canyon rim, 45, 52; stairway, 

59; steps, 149; trails, 84 
Harrington, John P., 15 
Harrington, M. R., 9 
Havasupai, the, 5 
Head, G., and A. Prieto, 41, 43 
hearth, 51, 62; firepits, 138, 150, 201; floor 

burns, 138, 201 
Heiken, Grant, 4, 15, 16, 134 
Hendron, J. W., 5, 8, 9, 11, 13, 15, 45, 

51, 62; excavation by, 63 
Herr, S., and R. Powers, 57 
Hewett, Edgar L., 1, 8, 14, 15, 17, 59, 77, 



296 CAVATE STRUCTURES 



80, 142, 145, 149, 182, 185, 190, 214, 

216 
Hill, James N., and W. Nicholas 

Trierweiler, 13, 72 
historic use of cavates, 7, 9, 71, 74 
hole, indeterminate, 108, 165, 174, 179, 

182, 201; holes and niches, 203. See 

also separately by hole type 
holes for pegs, 182 

holes in walls, multivariate analyses of, 174 
Hopi, the, 149, 214-16 
housekeeping, 94 

Hubbell, Lyndi, and Diane Traylor, 16 
Hyland, Justin R., 13, 62, 64, 116 

incised dado, 189, 233; illustrated, 153, 192 
incisions, narrow wall, 188, 201, 235 
indeterminate holes, 108, 165, 179, 182, 

201, 233; at the line of plastering, 174 
interior door, 122. See also doors 
irrigation, 28, 61 
isolated room (1-36), 37 

Jemez Mountains, 2, 15 
Jemez River, 2, 215 
Johnson, Chester, 11, 53 

Kapo Black ceramic, 71 

Kelley, V. C, E. H. Baltz, and R. A. 

Bailey, 16 
Kempe, David, 2, 4 
Kent, Kate P., 145, 149, 182, 215, 216 
Keres, the, 5, 14, 59, 217 
Kern, Willis F., and James R. Bland, 91, 92 
Kidder, A. V., 145 
kiva, 8, 12, 27-28, 51, 116, 206, 213-15, 

226; room categorized as, 13, 210, 211, 

227 
Kohler, Timothy A., 13, 45 
Kohler, Timothy A., and Angela R. Linse, 

17, 45, 52 

LA 211 (Tsankawi Pueblo), 16, 17, 19, 20, 

52, 71, 72, 74, 216 
LA 217 (Rainbow House), 45, 52, 63, 216 
LA 4997 (Saltbush Pueblo), 45, 63 
LA 50020. See Group M 



LA 50021. See Group A 

LA 50022. See Group I 

LA 50023. See Group F 

LA 50024. See Tsankawi 

LA 50909, ceramic dating for, 71 

LA 50972. See Group M 

LA 50973. See Group A 

LA 50974. See Group I 

LA 50975. See Group F 

LA 50976 (Tsankawi cavate group), 19, 53, 

216. See also Tsankawi 
LA 52333, excavation at, 13 
LA 60550 (Tyuonyi Annex), 216 
Laboratory of Anthropology, 9, 20; 

numbering by the, 90 
Lang, Richard W., 4, 65 
Lange, Charles H., and Carroll L. Riley, 5, 

6 
Lange, Charles H., Carroll L. Riley, and 

Elizabeth M. Lange, 6, 15 
large floor-level niche, 150 
Late Coalition Period, 216 
latilla hole, 162, 182, 201 
Lekson, Stephen H., 214 
Lent, Stephen C, 16 
levels (stories), 83, 98; multilevel rooms, 

94; Tsankawi, 112 
linguistics, 217 
lintels, 182 
Lister, Robert H., 11, 14, 28, 29, 37, 45, 

52, 53, 59, 86, 162 
Long House (Group D), 8, 11, 20, 216 

loom anchors, 144, 232, and upper 

loom supports, 182, illustrated, 192; 

kivas with loom features, 214-15, and 

comparisons, 215-16. See also loom 

features 
loom features, 201, 203, 206; loom anchors, 

144, 182, 214-16 
loom support, upper, 182, illustrated, 192. 

See also loom anchors 
Los Alamos Archaeological Society, 12 
Los Alamos Canyon, 52 
Los Alamos National Laboratory, 12 

McKenna, Peter J., 28, 61; ceramic analysis 
by, 64, 216 



INDEX 297 



McKenna, Peter J., and Robert P. Powers, 

64,65 
Magers, Pamela C, 214 
Mancos River, 4, 5 
mapping, 86, 90-91 
masonry, back wall, 84; comparisons of, 59; 

partial cavate, 84; presence and type of, 

97, 132, 226, 234, for functional 

analysis, 201; exterior masonry rooms, 

140; the use of, 94 
masonry and tuff wall, 132 
masonry wall, 132; masonry and natural 

wall, 234 
Mathien, F. Joan, Charlie R. Steen, and 

Craig D. Allen, 2, 15, 16, 17 
Maxon, James C, 9, 11-12, 196 
mealing complexes, adobe ramp in, 142; 

floor ridges and, 142; metate rest in, 

142; wall depression and, 185 
Mera, Harry P., 11 
metate rest, 142, 234; other mealing 

features, 142, 185 
Mills, Barbara M., 67, 185 
Mindeleff, Cosmos, 1, 4 
Mindeleff Cavate Site, 4 
Morley, Sylvanus G., 8-9 
Mortandad Canyon, 11 
Mortandad style of rock art, 196 

narrow wall incisions, 188, 235 

National Geographic, 13 

National Park Service, 14; database for the, 

93, 237, 240 
natural features, 97 
natural wall, 122, 126 
Navajo National Monument, 53 
Navawi site, 17 
New Cave, Arizona, 5 
NEWLANO, 90 

niche, cliff, 53, 84, 190. See also niches 
niche, deep modified, 13. See also niches 
niche, large floor-level, 150, illustrated, 

153. See also niches 
niches, as functional features, 201, and 

feature correlations with, 203; back wall 

niche, 151; cliff niche, 53, 84, 190; 

comparisons of types, 157; deep 



modified, 13; pairing of, 157, 
illustrated, 153; wall niche, 152, 157 

niche, wall, 152, 157; back wall niche, 
151. See also niches 

noncavates, back wall niches and, 151; back 
wall slots and, 157; connected with 
cavates, 84; data sets for, 83-84; 
exterior openings and, 122; features for, 
78, and distribution, 85; nonhuman 
users of, 99, 100, 227; viga holes in, 
161 

occupation, a sequence for, 62, 216; 
ceramic dating and, 67, 74; duration of, 
62, 206; Group F patterns of, 71; 
remodeling, 94, 157; reoccupation, 11, 
74, and historic use, 7, 9, 71, 74 

Old Caves, Arizona, 5 

Old World, cavates of the, 2, 4; in Turkey, 
4, 94, 134, 138 

Onstott, Thomas B., 75 note 2 

openings, as doors, 201; exterior, 116, 122; 
interior, 122; passages between 
chambers, 138. See also Doors 

Orcutt, Janet D., 67 

Otowi, 11, 16, 17 

overhang, 234; viga holes in cavate, 157 

Pajarito Archaeological Research Project 

(PARP), 13, 72 
Pajarito Plateau, described, 1, 3, 5-8, 15; 

past work on the, 2-13; settlement 

pattern for the, 216 
Panowski, Bruce, 90 
Panowski Holes, 61, 185, 235 
partial cavate, 84. See also noncavates 
partitions, vertical room, 185 
passage, 138, 234. See also doors 
Peckham, Stuart, 9, 145, 149, 213, 214 
Pecos Pueblo, 145 
pegs, holes for, 174, 182 
Peralta Canyon, 2 
petroglyphs, 12, 233, 234, 273; cliff face, 

272; defined, 190; Group A, 28; Group 

F, 29; Group I, 45; Tsankawi, 53 
photography, recording with, 86, 93, 244 
pictograph, defined, 190 



298 



CAVATE STRUCTURES 



pit complex, floor, 140 

pithouse, preceramic, 16 

pits, subfloor, 140 

plane shapes, 91, 232, 234 

planks, 142, 157 

plaster and niches, 203 

plastering, 94, 126-32, 206, 230; as 

functional feature, 201; colors of, 126, 

132, 233, 235; fine-line scratching in, 

190; floor features and, 203; 
indeterminate holes at the line of, 174, 

182; niches and, 150, 157; replastering, 

94; smoking and, 126. See also floor 

features 
population, aggregated, 17, 62; elevation 

and, 72 
possible latilla hole, 162, 182, 201 
possible upper loom support, 182 
postholes, 140, 235 
pot rests (floor depressions), 140 
Powell, John Wesley, Major, 1, 4, 6-7 
Powers, R., and T. Chadderdon, 55 
Powers, Robert P., 16, 17, 45, 61 
Preucel, Robert W., 2, 13, 17, 105, 215, 

216 
Prudden, T. Mitchell, 5 
Pueblo III, Late, 5 
Pueblo Canyon, 1 1 
Pueblo Revolt, 1 1 

pueblos, cavate relationship to surface, 216 
Puye, 13, 17; dating, 62, 63; described, 77; 

early visitors to, 6-9 

Rainbow House (LA 217), 45, 52, 63, 216 
remodeling, 94; combined chambers, 138; 

expanded chambers, 83; one from two 

chambers, 59; viga holes and, 157. See 

also occupation 
retaining wall, Group F, 29 
Riboud, Marc, 4, 94 
Rio Chama, 216, as Chama River, 15 
Rio Grande, 2, 4, 15, 16, 213 
Rio Grande (Ceramic) Series, 65, 66, 74 
Rio Grande Classic, 17, as Classic Period, 

62, 67, 71 
Rito de los Frijoles, 11, 15 
Robinson, William, 62, 63 



Robinson, William J., John W. Hannah, and 
Bruce G. Harrill, 62, 67 

rock art, as a functional feature, 201, with 
other features, 203; Chapman report on, 
8, 273; comparisons, 61, 192, 197, and 
discussed, 190; Crowder Study on, 197, 
267-72; Mortandad Style of, 12, 12-13, 
196; motifs for, 198, 233, and 
nomenclature, 199; recording rock art, 
85; rooms with notable, (Group M) 51, 
(Tsankawi) 53, 61, and others, 193, 
194,210,211,273 

roofing, beam features for, 201; viga holes, 
157; wall ledges, 185. See also ceilings 

Room A- 10, ceiling construction marks, 82; 
photograph, 32 

Room A-13, photographs, 30; rock art, 273 

Room A-47, features illustrated, 142, 153; 
pot rests, 140 

Room A-60, photographs, 31 

Room F-2, rock art, 273 

Room F-23, features illustrated, 193 

Room F-31, photograhs, 38 

Room F-37, features illustrated, 151 

Room F-38, rock art, 273 

Room F-47, step, 149 

Room 1-15, mentioned, 37 

Room 1-19, features, 211; kiva-like, 211; 
rock art, 273 

Room 1-22, photographs, 46 

Room 1-36, setting, 37 

Room M-10, expansion, 83 

Room M-13, kiva-like, 211; rock art, 194 

Room M-33, rock art, 51, 273 

Room M-40, features illustrated, 145 

Room M-44, features illustrated, 153 

Room M-59, features, 145, illustrated, 148, 
162 

Room M-60, masonry, 94; mealing activity, 
142, 185, illustrated, 146; rock art, 273, 
llustrated, 194 

Room of the Cacique, 5 

Room TS-20, kiva-like, 210 

Room TS-24, features illustrated, 189 

Room TS-25, features illustrated, 190 

Room TS-36, volume, 112 

Room TS-40, features illustrated, 193 



INDEX 299 



Room TS-53, photographs, 60 
Room TS-55, feature illustrated, 144 
Room TS-59, features, 112, 140, illustrated, 

148; kiva-like, 210; rock art, 196, 

illustrated, 192 
Room TS-64, features, 206; kiva-like, 210 
Room TS-65, features, 149 
Room TS-66, features, 112; kiva-like, 210 
Room 12, Group F, photographs, 39 
Room 15, Group F, features illustrated, 152 
rooms. See chambers; habitation rooms; 

kivas; noncavates; storage rooms; and 

see groups separately 
rooms up/rooms down, 82, 226 
Ross, C. S., R. L. Smith, and R. A. Bailey, 

2, 15 
routes to canyon rim, 45, 52; hand and/or 

toe holds, 84, 108, 189; stairway, 59; 

steps, 149; trails, 84 

SAS Institute, 93 

Saltbush Pueblo (LA 4997), 45, 63 

San Juan River, 4, 5 

San Lazaro Glaze-on-polychrome, 65 

Sandia Canyon, 52, 59 

Sankawi Black-on-cream, 52, 67, 71 

Santa Clara, the, 7 

Santa Fe, New Mexico, 93 

Santa Fe Black-on-white, 12, 13, 17, 62, 71 

Scaffold House, Navajo National 

Monument, 53 

Schaafsma, C, 87 

Schaafsma, Polly, 190, 191, 197 

seasonality, 215 

settlement pattern, 216. See also occupation 

Severy, Merle, 4 

shapes, 80, 232; plane shapes, 91, 234; 

solid shapes, 91, 233, 234 
Siegel, Sidney, 202 
slots, 157 

Smiley, Terah L., 67 
Smiley, Terah L., Stanley Stubbs, and 

Bryant Bannister, 62, 63 
Smith, Watson, 145, 213, 214 
smokehole, 182, 201 
smoking, 206 
Snake Village (Group E), 8 



Snead, J., 33 

Snead, J., and H. Newman, 35 

Snow, David H., 45, 63 

solid shapes, 91, 232, 233, 234 

Southwest, American, 4 

Southwest Museum, 9 

stability, 227, 245, 250, 253; table for, 100. 

See also condition/damage 
stairway, 59; hand and/or toe holds, 84, 

108, 189; route to canyon rim, 45, 52; 

steps, 149; trails, 84 
Steen, Charlie R., 12, 12-13, 14, 59, 134, 

196, 215 
steps, 149, 233. See also stairway 
Stevenson, James, 6 
storage rooms, defining, 13, 208, 211, 213, 

227, and comparing, 112, 116, 144; 

milling areas and, 144; plaster and 

smoking in, 134, 206 
stories (levels), multilevel rooms, 94; 

recording levels, 83, 98; Tsankawi, 1 12 
structural features, 108, 116, 122, 132 
Stuart, David E., and Rory P. Gauthier, 2, 

16,62 
subfloor pit, 140 
suites of cavates, 4 
Sun House (Group E), 8 
surface pueblos, cavate use and relationship 

to large, 216 
Sweetland, Bill, 28 

tent rocks, 29, 51 

terraces, 140 

Tesuque (ceramic) series, 74; Tesuque 

corrugated, 13 
Tewa, the, 5, 9, 14, 15, 52, 59, 217; rock 

art style and, 197 
Tewa polychrome, 71 
Thompson, Raymond H., 213 
trails, 84; hand and/or toe holds, 84, 108, 

189; routes to canyon rim, 45, 52; 

stairway, 59; steps, 149 
tree-ring dating, 9, 62-63; for Group M, 62 
Tsankawi, 11, 12, 14; ceramics and dating 

for, 62, 65, 67, 72; collections for, 85; 

comparisons, 60, 72, 112, and cultural 

separation of, 217; features for, 82, 84, 



300 CAVATE STRUCTURES 



116, 122, 132, 134, 138,140, 142, 145, 
149, 150, 157, 162, 185, 188, 190, 206; 
kivas at, 215; rock art at, 192, 196, 197, 
271, 272; setting of, 52-59, 61, 216, and 
room plans, 55, 57, 86, 91, 94, 108, 
112, 116,208 

Tsankawi Black-on-cream, 17 

Tsankawi Mesa, 17, 52 

Tsankawi Pueblo (LA 211), 16, 17, 19, 20, 
52, 71, 72, 74, 216 

Tsankawi Pumice Bed, 16, 53 

Tshirege Cave Site, 12, 17, 63 

Tsiping Ruin, 2 

tuff type, 82, 97, 99, 227, 233, 235; 

Bandelier Tuff, 3, 9, 15, 16; patina and 
smoking on tuff, 134 

Turkey, comparisons to Cappadocia, 4, 94, 
134, 138 

Turkey Tank Cave, Arizona, 5 

Turney, John F., 9-11, 51, 71 

Tyuonyi Annex (LA 60550), 216 

Tyuonyi Pueblo, 5, 17, 29, 63, 216 

United States Geological Survey (USGS), 20 
Universal Transverse Mercator (UTM) grid, 

90-91 
University of California, 13 
University of New Mexico, 11 
Unshagi site, 215 

Valle Grande Caldera, 3, as Valles Caldera, 

15 
Van Zandt, Tineke R., 75 note 2 
vents, 185, 203; as functional features, 201 



Verde Valley, Arizona, 4 

vertical ceiling hole, 61, 185, 235 

video recording, 86, 88, 244 

viga holes, 107, 157, and beam seats, 165; 

as beam features, 201. See also beam 

features 
visitation, 101; at Group A, 28. See also 

damage 
volumes and areas, 91, 239; volume and 

niches, 203 

wall depression, 185, 234 

wall features, 150 

wall ledges, 185, 234, as beam features, 

201 
wall niche, 152, 157 
wall, masonry, 132; masonry and tuff, 132; 

natural wall, 122, 126 
walls, as functional features, 201; back 

walls, 84, 122, 151, 157, 161; 

depression in, 185; features for, 150; 

ledges in, 185, 201; masonry of, 132, 

and natural, 122, 126; niches in, 

152, 157; recording, 230 
Washington State University, 13 
weaving, 215. See also loom features 
White, G. E., 1, 134 
White Rock, New Mexico, 12, 15 
Whittaker, John C, 5 
Wiyo Black-on-white, 12, 71 
wood, features made of, 144 

Yapashi site, 72 



oU.S. GOVERNMENT PRINTING OFFICE: 1995-676-953/25132 



Clemson University 



3 1604 01 



357 938 




FEDERAL 
PUBLICATION 



DATE DUE 







































































DEMCO, INC. 38-2931 



■ 



^B